Bioelectrochemistry: fundamentals, experimental techniques and applications 9780470843642, 0470843640, 0470753846, 9780470753842

Bioelectrochemistry: Fundamentals, Experimental Techniques and Application , covers the fundamental aspects of the chemi

284 68 7MB

English Pages 488 Year 2008

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Halftitle......Page 2
Inside Cover......Page 3
Copyright......Page 4
Contents......Page 5
List of Contributors......Page 12
Preface......Page 14
1 Bioenergetics and Biological Electron Transport......Page 16
2 Electrochemistry of Redox Enzymes......Page 53
3 Biological Membranes and Membrane Mimics......Page 100
4 NAD(P)-Based Biosensors......Page 170
5 Glucose Biosensors......Page 212
6 Phenolic Biosensors......Page 231
7 Whole-Cell Biosensors......Page 261
8 Modelling Biosensor Responses......Page 279
9 Bioelectrosynthesis–Electrolysis and Electrodialysis......Page 338
10 Biofuel Cells......Page 370
11 Electrochemical Immunoassays......Page 387
12 Electrochemical DNA Assays......Page 421
13 In Vivo Applications: Glucose Monitoring, Fuel Cells......Page 453
Index......Page 477
Recommend Papers

Bioelectrochemistry: fundamentals, experimental techniques and applications
 9780470843642, 0470843640, 0470753846, 9780470753842

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Bioelectrochemistry

Bioelectrochemistry: Fundamentals, Experimental Techniques and Applications Edited by Philip Bartlett # 2008 John Wiley & Sons, Ltd. ISBN: 978-0-470-84364-2

Bioelectrochemistry Fundamentals, Experimental Techniques and Applications

Edited by P. N. Bartlett University of Southampton, UK

Copyright Ó 2008

John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England Telephone

(þ 44) 1243 779777

Email (for orders and customer service enquiries): [email protected] Visit our Home Page on www.wileyeurope.com or www.wiley.com All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W1T 4LP, UK, without the permission in writing of the Publisher. Requests to the Publisher should be addressed to the Permissions Department, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or emailed to [email protected], or faxed to (þ 44) 1243 770620. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The Publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the Publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. The Publisher and the Author make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of fitness for a particular purpose. The advice and strategies contained herein may not be suitable for every situation. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. The fact that an organization or Website is referred to in this work as a citation and/or a potential source of further information does not mean that fee author or the publisher endorses the information the organization or Website may provide or recommendations it may make. Further, readers should be aware that Internet Websites listed in this work may have changed or disappeared between when this work was written and when it is read. No warranty may be created or extended by any promotional statements for this work. Neither the Publisher nor the Author shall be liable for any damages arising herefrom. Other Wiley Editorial Offices John Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USA Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA Wiley-VCH Verlag GmbH, Boschstr. 12, D-69469 Weinheim, Germany John Wiley & Sons Australia Ltd, 42 McDougall Street, Milton, Queensland 4064, Australia John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809 John Wiley & Sons Ltd, 6045 Freemont Blvd, Mississauga, Ontario L5R 4J3, Canada Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. Library of Congress Cataloging-in-Publication Data Bioelectrochemistry : fundamentals, experimental techniques, and applications / editor-in-chief, P.N Bartlett ; editorial advisory board, T. Cass ... [et al.]. p. ; cm. Includes bibliographical references and index. ISBN 978-0-470-84364-2 (cloth : alk. paper) 1. Bioelectrochemistry. I. Bartlett, P. N. [DNLM: 1. Energy Metabolism–physiology. 2. Bioelectric Energy Sources. 3. Biological Transport. 4. Biosensing Techniques. 5. Electron Transport. 6. Models, Biological. QU 125 B6144 2008] QP517.B53B5435 2008 572’.437–dc22 2007046620 British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN 978-0470-843642 Typeset in 9/11 pt Times by Thomson Digital, India Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire

Contents

List of Contributors Preface 1 Bioenergetics and Biological Electron Transport Philip N. Bartlett 1.1 Introduction 1.2 Biological Cells 1.3 Chemiosmosis 1.3.1 The Proton Motive Force 1.3.2 The Synthesis of ATP 1.4 Electron Transport Chains 1.4.1 The Mitochondrion 1.4.2 The NADH–CoQ Reductase Complex 1.4.3 The Succinate–CoQ Reductase Complex 1.4.4 The CoQH2–Cyt c Reductase Complex 1.4.5 The Cyt c Oxidase Complex 1.4.6 Electron Transport Chains in Bacteria 1.4.7 Electron Transfer in Photosynthesis 1.4.8 Photosystem II 1.4.9 Cytochrome bf Complex 1.4.10 Photosystem I 1.4.11 Bacterial Photosynthesis 1.5 Redox Components 1.5.1 Quinones 1.5.2 Flavins 1.5.3 NAD(P)H 1.5.4 Hemes 1.5.5 Iron–Sulfur Clusters 1.5.6 Copper Centres 1.6 Governing Principles 1.6.1 Spatial Separation 1.6.2 Energetics: Redox Potentials 1.6.3 Kinetics: Electron Transfer Rate Constants 1.6.4 Size of Proteins 1.6.5 One-Electron and Two-Electron Couples 1.7 ATP Synthase 1.8 Conclusion References

xiii xv 1 1 2 3 3 5 6 6 8 9 9 10 12 12 15 16 17 18 19 19 20 20 21 22 22 22 23 25 27 31 32 32 34 34

vi

Contents

2 Electrochemistry of Redox Enzymes James F. Rusling, Bingquan Wang and Sei-eok Yun 2.1

Introduction 2.1.1 Historical Perspective 2.1.2 Examples of Soluble Mediators 2.1.3 Development of Protein-Film Voltammetry and Direct Enzyme Electrochemistry 2.2 Mediated Enzyme Electrochemistry 2.2.1 Electron Mediation 2.2.2 Wiring with Redox Metallopolymer Hydrogels 2.2.3 Wiring with Conducting Polymers 2.2.4 NAD(P) þ /NAD(P)H Dependent Enzymes 2.2.5 Regeneration of NAD(P)H from NAD(P) þ 2.2.6 Regeneration of NAD(P) þ from NAD(P)H 2.3 Direct Electron Transfer between Electrodes and Enzymes 2.3.1 Enzymes in Solution 2.3.2 Enzyme-Film Voltammetry: Basic Theory 2.3.3 Adsorbed and Coadsorbed Enzyme Monolayers 2.3.4 Self-Assembled Monolayers and Covalently Attached Enzymes 2.3.5 Enzymes on Carbon Nanotube Electrodes 2.3.6 Enzymes in Lipid Bilayer Films 2.3.7 Polyion Films and Layer-by-Layer Methods 2.4 Outlook for the Future Acknowledgements References 3 Biological Membranes and Membrane Mimics Tibor Hianik 3.1 Introduction 3.2 Membrane Structure and Composition 3.2.1 Membrane Structure 3.2.2 Membrane Lipids 3.2.3 Membrane Proteins 3.3 Models of Membrane Structure 3.3.1 Lipid Monolayers 3.3.2 Bilayer Lipid Membranes (BLM) 3.3.3 Supported Bilayer Lipid Membranes 3.3.4 Liposomes 3.4 Ordering, Conformation and Molecular Dynamics of Lipid Bilayers 3.4.1 Structural Parameters of Lipid Bilayers Measured by X-ray Diffraction 3.4.2 Interactions between Bilayers 3.4.3 Dynamics and Order Parameters of Bilayers Determined by EPR and NMR Spectroscopy and by Optical Spectroscopy Methods 3.5 Phase Transitions of Lipid Bilayers 3.5.1 Lyotropic and Thermotropic Transitions 3.5.2 Thermodynamics of Phase Transitions 3.5.3 Trans–Gauche Isomerization 3.5.4 Order Parameter 3.5.5 Cooperativity of Transition 3.5.6 Theory of Phase Transitions

39 39 39 40 41 42 42 43 49 52 53 55 57 57 58 63 65 67 68 71 75 76 76 87 87 88 88 90 94 95 95 100 106 107 107 108 109 111 116 116 117 117 118 119 119

Contents

3.6

Mechanical Properties of Lipid Bilayers 3.6.1 Anisotropy of Mechanical Properties of Lipid Bilayers 3.6.2 The Model of an Elastic Bilayer 3.6.3 Mechanical Properties of Lipid Bilayers and Protein–Lipid Interactions 3.7 Membrane Potentials 3.7.1 Diffusion Potential 3.7.2 Electrostatic Potentials 3.7.3 Methods of Surface Potential Measurement 3.8 Dielectric Relaxation 3.8.1 The Basic Principles of the Measurement of Dielectric Relaxation 3.8.2 Application of the Method of Dielectric Relaxation to BLMs and sBLMs 3.9 Transport Through Membranes 3.9.1 Passive Diffusion 3.9.2 Facilitated Diffusion of Charged Species Across Membranes 3.9.3 Mechanisms of Ionic Transport 3.9.4 Active Transport Systems 3.10 Membrane Receptors and Cell Signaling 3.10.1 Physical Reception 3.10.2 Principles of Hormonal Reception 3.10.3 Taste and Smell Reception 3.11 Lipid-Film Coated Electrodes 3.11.1 Modification of Lipid-Film Coated Electrodes by Functional Macromolecules 3.11.2 Bioelectrochemical and Analytical Applications of Lipid Coated Electrodes Acknowledgements References 4 NAD(P)-Based Biosensors L. Gorton and P. N. Bartlett 4.1 4.2 4.3

4.4

4.5

4.6

4.7

Introduction Electrochemistry of NAD(P) þ /NAD(P)H Direct Electrochemical Oxidation of NAD(P)H 4.3.1 General Observations 4.3.2 Effect of Adsorption 4.3.3 Mechanism and Kinetics Soluble Cofactors 4.4.1 Implications of the Low Eo0 Value for Practical Applications 4.4.2 Special Prerequisites for Biosensors Based on NAD(P)-Dependent Dehydrogenases Mediators for Electrocatalytic NAD(P)H Oxidation 4.5.1 Other Mediating Functionalities and Metal Coated Electrodes 4.5.2 Electropolymerisation 4.5.3 Carbon Paste 4.5.4 Gold Nanoparticles Construction of Biosensors from NAD(P)H-Dependent Dehydrogenases 4.6.1 Entrapment Behind a Membrane 4.6.2 Covalent Attachment to a Nylon Net or Membrane 4.6.3 Cross-Linking 4.6.4 Entrapment in a Polymer Film 4.6.5 Carbon Paste 4.6.6 Self-Assembled Monolayers Conclusions

vii

120 120 122 123 127 127 127 129 132 132 132 133 134 135 137 139 139 140 140 142 142 142 143 148 148 157 157 159 160 160 160 161 163 164 165 165 181 183 183 183 183 183 184 184 184 184 185 185

viii

Contents

Acknowledgements References 5 Glucose Biosensors Josep M. Montornes, Mark S. Vreeke and Ioanis Katakis 5.1 5.2

185 185 199

Introduction to Glucose Sensors Biosensors 5.2.1 Types of Sensors 5.2.2 Transduction Mode 5.3 Application Areas 5.3.1 Clinical 5.3.2 Food and Fermentation 5.4 Design Requirements 5.4.1 Disposable Glucose Sensor 5.4.2 Continuous Glucose Sensor 5.4.3 Implantable Glucose Sensor 5.5 Biosensor Construction 5.5.1 Artificial Mediators 5.5.2 Immobilization of GOx 5.5.3 Inner and Outer Membrane Function 5.6 From Product Design Requirements to Performance 5.6.1 Design Exercise for the Disposable Glucose Sensor 5.7 Conclusions Acknowledgement References

199 200 200 200 202 202 202 202 202 203 204 204 204 205 206 207 207 212 213 213

6 Phenolic Biosensors Ulla Wollenberger, Fred Lisdat, Andreas Rose and Katrin Streffer

219

6.1 6.2

Introduction Enzymes Used for Phenol Biosensors 6.2.1 Phenol Oxidation by Water-Producing Oxidases and Oxygenases 6.2.2 Reducing Enzymes 6.3 Design of Phenol Biosensors 6.3.1 Oxygen Consumption 6.3.2 Bioelectrocatalysis Based on Phenol-Oxidizing Enzymes 6.3.3 Electrocatalytic Sensors Based on Quinone-Reducing Enzymes 6.4 Applications 6.4.1 Waste Water Treatment 6.4.2 Sensitive Label Detectors 6.4.3 Catecholamines 6.5 Summary and Conclusions Acknowledgements References 7 Whole-Cell Biosensors H. Shiku, K. Nagamine, T. Kaya, T. Yasukawa and T. Matsue 7.1 7.2

7.3

Introduction Whole-Cell Biosensors Probing Cellular Functions 7.2.1 Redox Reactions in Whole Cells 7.2.2 Responses on a Microbial Chip with Collagen Gel Whole-Cell Microdevices Fabricated Using Bio-MEMS Technologies 7.3.1 Potentiometric Devices: LAPS and ISFET

219 220 220 225 227 227 229 236 237 237 238 239 241 241 241 249 249 252 252 253 255 255

Contents

7.3.2 Other Electrochemical Devices: Amperometric and Impedance Sensors 7.3.3 Improvement of Cell Culture within Microenvironments 7.4 Genetically Engineered Whole-Cell Microdevices 7.4.1 Sensors with Gene-Modified Bacteria 7.4.2 Transcriptional Responses on a Microbial Chip with Collagen Gel 7.4.3 Cellular Devices for High-Throughput Screening 7.4.4 Microdevice for On-Chip Transfection and On-Chip Transformation 7.5 Conclusions References 8 Modelling Biosensor Responses P. N. Bartlett, C. S. Toh, E. J. Calvo and V. Flexer 8.1 8.2

Introduction Enzyme Kinetics 8.2.1 Equilibrium and Steady-State 8.2.2 Analysis of Enzyme Kinetic Data 8.2.3 The Significance of KMS for Biosensor Applications 8.3 Modelling Enzyme Electrodes 8.3.1 The Flux Diagram for the Membrane|Enzyme|Electrode 8.3.2 Solving the Coupled Diffusion/Reaction Problem for the Membrane Enzyme|Electrode 8.3.3 Deriving a Complete Kinetic Model 8.3.4 Experimental Verification of Approximate Analytical Kinetic Models 8.4 Numerical Simulation Methods 8.4.1 Explicit Numerical Methods 8.4.2 The Crank–Nicholson Method 8.4.3 Other Simulation Methods 8.5 Modelling Redox Mediated Enzyme Electrodes 8.5.1 Steady-State Kinetics 8.5.2 Homogeneous Mediated Enzyme Electrode 8.6 Modelling Homogeneous Enzyme with Attached Redox Mediator 8.7 Non-Steady-State Techniques for Homogeneous Enzyme Systems 8.7.1 Extraction of Kinetic Parameters 8.8 The Heterogeneous Mediated Mechanism 8.8.1 Enzyme Monolayers with Soluble Redox Mediator 8.8.2 Enzyme Multilayers 8.9 Conclusions Acknowledgements References 9 Bioelectrosynthesis–Electrolysis and Electrodialysis Derek Pletcher 9.1 9.2

9.3

Introduction Setting the Scene 9.2.1 Electrolytic Production of Organic Compounds 9.2.2 Technological Factors 9.2.3 Enzymes in Organic Synthesis 9.2.4 Combining Enzyme Chemistry and Electrosynthesis Mechanisms of and Approaches to Bioelectrosynthesis 9.3.1 Homogeneous Systems 9.3.2 Electrode Coatings

ix

255 256 257 257 259 260 261 262 263 267 267 268 268 268 269 270 273 276 278 280 280 281 282 283 286 287 287 294 295 300 302 304 306 317 317 317 327 327 327 327 329 333 334 335 335 339

x

Contents

9.4

Examples of Syntheses 9.4.1 The Oxidation of Alcohols and Diols 9.4.2 The Oxidation of 4-Alkylphenols 9.4.3 The Synthesis of Dihydroxyacetone Phosphate 9.4.4 The Site Specific Oxidation of Sugars 9.4.5 Hydroxylation of Unactivated C–H Bonds 9.4.6 Reduction of Carbonyl Compounds 9.4.7 Hydrogenation 9.4.8 Conclusions 9.5 Electrodialysis 9.6 Conclusions References 10 Biofuel Cells G. Tayhas R. Palmore 10.1 10.2

Introduction Fundamentals of Fuel Cells 10.2.1 How Fuel Cells Work 10.2.2 Equations that Govern the Performance of a Fuel Cell 10.3 Economics of Conventional Fuel Cells and Biofuel Cells 10.4 Biofuel Cells that use Micro-Organisms as the Catalytic Element 10.5 Biofuel Cells that use Oxidoreductases as the Catalytic Element 10.6 Future Directions in Biofuel Cell Research Acknowledgements References

11 Electrochemical Immunoassays Julia Yakovleva and Jenny Emneus 11.1 11.2

Introduction Basic Concepts in Immunoassay 11.2.1 Antibodies: Structure, Properties and Production 11.2.2 Classification of Immunoassays 11.3 Electrochemical Immunoassays (ECIA) 11.3.1 Labels in ECIAs 11.3.2 Electrochemical Detection Principles 11.4 Different Assay Formats and Applications 11.4.1 Heterogeneous ECIA 11.4.2 Homogeneous ECIAs 11.4.3 Electrochemical Immunosensors 11.5 Future Perspectives and Recent Trends References 12 Electrochemical DNA Assays Ana Maria Oliveira-Brett 12.1 12.2

12.3

Introduction Electrochemistry of DNA 12.2.1 Reduction 12.2.2 Oxidation Adsorption of DNA at Electrode Surfaces 12.3.1 Guanine Adsorbates 12.3.2 Adsorbed dsDNA and ssDNA

340 340 342 344 344 345 346 348 348 350 354 355 359 359 359 359 360 363 363 366 369 371 372 377 378 378 378 379 380 380 388 394 394 399 401 402 403 411 411 413 413 414 416 416 417

Contents

12.4

Electrochemistry for Sensing/Probing DNA Interactions 12.4.1 Biomarkers for DNA Damage 12.4.2 DNA–Metal Interactions 12.4.3 DNA–Drug Interactions 12.5 DNA Electrochemical Biosensors 12.5.1 In Situ DNA Oxidative Damage 12.5.2 DNA Hybridisation 12.5.3 DNA Microarrays 12.6 Conclusions Acknowledgements References 13 In Vivo Applications: Glucose Monitoring, Fuel Cells P. Vadgama and M. Schoenleber 13.1 13.2

Introduction Biocompatibility 13.2.1 General Concepts 13.2.2 Protein Constituents 13.2.3 Blood 13.2.4 Tissue 13.3 Materials Interfacing Strategy 13.4 Implanted Glucose Biosensors 13.4.1 Electrode Designs for Tissue Monitoring 13.4.2 Electrode Designs for Blood Monitoring 13.4.3 Early Phase Tissue Response 13.4.4 Open Microflow 13.4.5 Bioelectrochemical Fuel Cells 13.5 Conclusions References Index

xi

422 423 424 426 430 430 431 435 436 436 436 443 443 444 444 446 447 449 450 451 451 456 457 458 460 462 463 467

List of Contributors

P. N. Bartlett Department of Chemistry, University of Southampton, Southampton, SO17 1BJ, UK E. J. Calvo INQUIMAE, Departamento de Quı´mica Inorganica, Analı´tica, y Quı´mica Fı´sica, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Pabello´n 2, Ciudad Universitaria, AR-1428 Buenos Aires, Argentina Jenny Emneus Department of Analytical Chemistry, Center for Chemistry and Chemical Engineering, Lund University, SE-22100, Lund, Sweden V. Flexer INQUIMAE, Departamento de Quı´mica Inorganica, Analı´tica, y Quı´mica Fı´sica, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Pabello´n 2, Ciudad Universitaria, AR-1428 Buenos Aires, Argentina L. Gorton Department of Analytical Chemistry, Lund University, PO Box 124, SE-221 00 Lund, Sweden Tibor Hianik Department of Biophysics and Chemical Physics, Comenius University, Bratislava, Slovakia Ioanis Katakis Bioengineering and Bioelectrochemistry Group, Chemical Engineering Department, Rovira i Virgili

University, Avinguda Paı¨sos Catalans, 26, 43007 Tarragona, Catalonia, Spain T. Kaya Graduate School of Environmental Studies & Graduate School of Engineering, Tohoku University, Aramaki 7, Sendai 980-8579, Japan Fred Lisdat University of Applied Sciences, Bahnhofstrasse, 15745 Wildau, Germany T. Matsue Graduate School of Environmental Studies & Graduate School of Engineering, Tohoku University, Aramaki 7, Sendai 980-8579, Japan Josep M. Montornes DINAMIC Technology Innovation Centre, Avinguda Paı¨sos Catalans, 18, 43007 Tarragona, Catalonia, Spain K. Nagamine Graduate School of Environmental Studies & Graduate School of Engineering, Tohoku University, Aramaki 7, Sendai 980-8579, Japan Ana Maria Oliveira-Brett Departamento de Quı´mica, Faculdade de Ci^encias e Tecnologia, Universidade de Coimbra, 3004-535 Coimbra, Portugal

xiv

List of Contributors

G. T. R. Palmore Division of Engineering and Division of Biology and Medicine, Brown University, Providence, Rhode Island, USA

P. Vadgama IRC in Biomedical Materials, Queen Mary, University of London, Mile End Road, London, El 4NS, UK

Derek Pletcher Department of Chemistry, The University of Southampton, Southampton SO17 1BJ, UK

Mark S. Vreeke Rational Systems, 8 Greenway Plaza, Houston, TX, USA

Andreas Rose University of Potsdam, Institute of Biology and Biochemistry, Liebknechtstrasse 24-25/H25, 14476 Golm, Germany

Bingquan Wang Department of Chemistry, University of Connecticut, Storrs, CT 06269-3060, USA

James F. Rusling Department of Chemistry and Department of Pharmacology, University of Connecticut, Storrs, CT 06269-3060, USA

Ulla Wollenberger University of Potsdam, Institute of Biology and Biochemistry Department Analytical Biochemistry, Liebknechtstrasse 24-25/ H25, D-14476 Golm, Germany

M. Schoenleber IRC in Biomedical Materials, Queen Mary, University of London, Mile End Road, London, El 4NS, UK H. Shiku Graduate School of Environmental Studies & Graudate School of Engineering, Tohoku University, Aramaki 7, Sendai 980-8579, Japan Katrin Streffer University of Potsdam, Institute of Biology and Biochemistry, Liebknechtstrasse 24-25/H25, 14476 Golm, Germany C. S. Toh Department of Chemistry, National University of Singapore, 3 Science Drive 3, Singapore 117543

Julia Yakovleva Department of Analytical Chemistry, Center for Chemistry and Chemical Engineering, Lund University, SE-22100, Lund, Sweden T. Yasukawa Graduate School of Environmental Studies & Graduate School of Engineering, Tohoku University, Aramaki 7, Sendai 980-8579, Japan Sei-eok Yun Department of Food Science and Technology, Chonbuk National University, Jeonju, Republic of Korea

Preface

Electron transfer reactions between molecules at interfaces play a central role in all living systems. In respiration and in photosynthesis, sequential electron transfer at and across biological membranes is the central process in the generation of the proton motive force, an essentially electrochemical phenomenon, which is used to drive the thermodynamically uphill synthesis of adenosine triphosphate (ATP) from inorganic phosphate and adenosine diphosphate (ADP). ATP is the essential energy carrier which is used in a wide range of biological processes where the energy produced by the hydrolysis of ATP back to ADP and inorganic phosphate is used to drive many different and essential reactions in living cells. Thus electrochemical processes, that is electron transfer at interfaces and the build up of potential differences at and across interfaces, are central processes in life. Bioelectrochemistry is the study and application of biological electron transfer processes. Over the last 25 years there have been enormous advances in bioelectrochemistry and in the study of electrochemical reactions of biological molecules both large, such as redox proteins and redox enzymes, and small, such as NADH and quinones. Over this time we have learnt some of the important factors which control the interaction between biological redox partners. We have learnt how to apply this knowledge and to start to design electrode surfaces, through deliberate chemical modification, so that the biological molecules will interact in a productive way with the electrode surface and facilitate efficient electron transfer. Over the same period, significant parallel developments in physical electrochemistry have meant that the tools and techniques, such as in situ infrared spectroscopy, SERS, EQCM, STM and AFM, now exist to study the electrode solution interface in situ at the molecular level. These techniques are now being used to characterise chemically modified electrode surfaces and to study their interaction with biological molecules. This has led to increasing sophistication in the approaches to the assembly of molecules on electrode surfaces, including methods for the immobilisation and orientation of enzymes at the electrode surface and the use of techniques from molecular biology to produce mutant enzymes modified to facilitate electrochemical applications. At the same time electrochemical approaches have been developed and successfully applied to study the electron transfer behaviour of complex multicentre redox enzymes and to investigate the gating of electron transfer, thus using electrochemical techniques to address fundamental biological questions. Bioelectrochemistry has many applications in practical devices such as biosensors, where the different and highly successful electrochemical whole blood glucose sensors developed for use by diabetics are a notable example, biofuel cells, an area of significant current interest, and sensors based on applications of biological membranes. Bioelectrochemistry is a field of research which continues to expand rapidly. It is an area which brings together scientists from a variety of disciplines including chemistry, physics and biology. The aim of this book is to provide a modern view of the field which will be accessible to graduate students and final year undergraduate students in chemistry and biochemistry as well as researchers in related disciplines including biology, physics, physiology and pharmacology. The work covers the fundamental aspects of the chemistry, physics and biology which underline the subject area. It describers some of the different experimental techniques that can be used to study bioelectrochemical problems and it describes various applications of bioelectrochemistry including amperometric biosensors, immunoassays, electrochemistry of DNA, biofuel cells, whole cell biosensors, in vivo applications and bioelectrosynthesis.

xvi

Preface

I would like to thank all of my coauthors for their enthusaiasm and support in producing this book and for their patience. I hope that you, the reader, will find this book useful and instructive, but more than that I hope that it will inspire you to undertake further research in bioelectrochemistry and to directly contribute to this exciting and important interdisciplinary field. Philip Bartlett University of Southampton

1

Bioenergetics and Biological Electron Transport Philip N. Bartlett School of Chemistry, University of Southampton, Southampton, SO17 1BJ, UK

1.1 INTRODUCTION Electron transfer reactions play a central role in all biological systems because they are essential to the processes by which biological cells capture and use energy. These electron transfer reactions occur in highly organised ways, in electron transport chains in which electron transfer occurs in an ordered way between specific components, and these electron transfer reactions occur at interfaces. In this chapter we will explore the principles behind the organisation and operation of these electron transfer chains from an electrochemical perspective. We will examine the guiding physical principles which govern the efficient operation of biological electron transfer. As we will see, several guiding principles emerge: the tuning of redox potentials for different components in the electron transfer chain to optimise energy efficiency, the control of distance between redox centres to control the kinetics of electron transfer and to achieve specificity, the role of an insulating lipid bilayer to separate charge and store electrochemical energy. Such a study is informative, not only because it tells us about the structure, organisation and function of biological systems, but also because we can learn useful lessons from the study of biological electron transfer systems which have evolved over millions of years which we can use to guide our design of electrochemical systems. For example, electrocatalysis of the four-electron reduction of oxygen to water at neutral pH remains a key barrier to the development of efficient polymer electrolyte membrane (PEM) fuel cells. This same reaction is an important component in the

mitochondrial electron transport chain where it is achieved using non-noble metal catalytic sites. A detailed understanding of this biological reaction may give clues to the design of new electrocatalysts for fuel cells. Similarly an understanding of the organisation, light harvesting and electron transfer reactions in the photosynthetic systems in plants and bacteria can inform our design of artificial photosynthetic systems for solar energy conversion. Closer to home, an understanding of the principles which govern efficient biological electron transfer is essential if we wish to exploit biological electron transfer components, such as oxidoreductase enzymes, NADH-dependent dehydrogenases or redox proteins, in biosensors, biofuel cells or bioelectrosynthesis. This chapter is conceived as a general introduction to biological electron transfer processes for those with little or no prior knowledge of the subject, but with a background in chemistry or electrochemistry. As such it should serve as an introduction to the more specific material to be found in the chapters which follow. At the same time I have tried to emphasise the underlying principles, as seen from an electrochemical perspective, and to bring out similarities rather than to emphasise the differences in detail between the different electron transport chains. Such an interest in the organisation and principles which guide biological electron transfer is directly relevant to current interest in integrated chemical systems [1]. Broadly this chapter is organised as follows. We begin with a very simple description of the different types of

Bioelectrochemistry: Fundamentals, Experimental Techniques and Applications Edited by Philip Bartlett # 2008 John Wiley & Sons, Ltd. ISBN: 978-0-470-84364-2

2

Bioelectrochemistry

biological cell, bacterial, plant and animal, their internal organisation and the different structures within them. We then consider the structure of biological cells from an electrochemical perspective focusing on the processes of energy transduction and utilisation. This is followed by a more detailed description of the electron transfer chains in the mitochondria and in the photosynthetic membrane and the different redox centres that make up the electron transport chains in these systems. We then describe the governing principles which emerge as the important features in all of these electron transfer processes, before concluding with a discussion of the way these processes are used to drive the thermodynamically uphill synthesis of adenosine triphosphate (ATP). 1.2 BIOLOGICAL CELLS All living matter is made up of cells and these cells share many common features in terms of their structure and the chemical components which make up the cell. The different types of living cell, from the simplest bacteria to complex plant and mammalian cells, carry out many of their fundamental processes in the same way. Thus the production of chemical energy by conversion of glucose to carbon dioxide is carried out in a similar way across all biological cells. This similarity reflects the common ancestry of all living cells and the process of evolution. In this section we focus on the internal structure of the different types of biological cell. For a more detailed

discussion of biological cells the reader is directed to modern biochemistry or cell biology texts such as that by Lodish and colleagues which provides a beautifully illustrated account of the subject [2]. Biological cells can be divided into two classes: eukaryotic and prokaryotic (Figure 1.1). The eukaryotes include all plants and animals as well as many single cell organisms. The prokaryotes have a simpler cell structure and include all bacteria; they are further divided into the eubacteria and archaebacteria. In the prokaryotic cell there is a single plasma membrane, a phospholipid bilayer, which separates the inside of the cell from the outer world, although in some cases there can also be simple internal photosynthetic membranes. In contrast, in eukaryotes the inner space within the cell is further divided into a number of additional structures called organelles. These are specialised structures surrounded by their own plasma membranes. Thus, within the eukaryotic cell we find specialised structures, such as the nucleus, which contains the cell’s DNA and which directs the synthesis on RNA within the cell, peroxisomes, which metabolise hydrogen peroxide, mitochondria, where ATP is generated by oxidation of small molecules, and, in plants, chloroplasts, where light is captured. It is the last two of these, the mitochondria and the chloroplasts which are of most interest to us here since both are central to biological energy transduction and both function electrochemically. These same functions of energy transduction occur in prokaryotes,

All Biological Cells

Prokaryotes All Bacteria

Eukaryotes

all plants (simple ferns to complex plants) all animals (sponges, insects, mammals) single celled organisms true algae amobae fungi moulds

eubacteria – the majority blue-green algae

archaebacteria – less familiar, found in unusual environments e.g. methanogens, halophiles, thermacidophiles

Figure 1.1 The general classification of biological cells.

Bioenergetics and Biological Electron Transport

but in this case they are associated with the outer cell membrane. Below we give a more detailed account of the electron transfer processes which occur in the mitochondrion and chloroplast, but for now we concentrate on the essential common features. Both processes, the oxidation of small molecules to generate energy in the mitochondrion and the capture of light and its transduction into energy that the cell can use in the chloroplast, occur across energy transducing membranes. In both cases the final product is ATP (see below), a high energy species that is used elsewhere in the cell to drive catabolism (the synthesis of molecules within the cell) and other living processes. An essential feature of the phospholipid bilayers which make up the plasma membranes of the cell and the different organelles within the eukaryotic cell is that although they are permeable to gases such as oxygen and carbon dioxide, they are impermeable to larger molecules such as amino acids or sugars and they are impermeable to ions such as Hþ, Kþ or Cl. This allows the cell to control the composition of the solution on the two sides of the membrane separately, a process which is achieved by the presence of specific transmembrane proteins, or permeases, within the cell membrane, which control transport of molecules and ions across the membrane. The energy transducing membranes of eukaryotes and prokaryotes, that is the plasma membrane of simple prokaryotic cells such as bacteria and blue-green algae, the inner membrane of mitochondria and the thylakoid membrane of chloroplasts in eukaryotes, share many common features. All of these membranes have two distinct protein assemblies: the ATPase at which ADP is converted to adenosine triphosphate (ATP) and the energy source electron transport chain which provides the thermodynamic driving force to the synthesis of ATP. These two processes are linked by the directed flow of electrons and protons across the membrane in order to establish an electrochemical potential which is used to drive ATP synthesis. This chemiosmotic model of biological energy transduction, which is essentially an electrochemical model, was first described by Mitchell in 1961 and was recognised by the award of the Nobel prize for chemistry in 1978 [3,4]. 1.3 CHEMIOSMOSIS The key concept in the chemiosmotic theory is that the synthesis of ATP is linked to the energy source electron transfer chain through the transmembrane proton motive force that is set up. This proton motive force is made up of

3

a contribution from the proton concentration gradient across the membrane, as well as the potential difference across the membrane. Figure 1.2 shows a simplified picture for the mitochondrial membrane and the thylakoid membrane of the chloroplast. In the mitochondria and aerobic bacteria, energy from the oxidation of carbon compounds, such as glucose, is used to pump protons across the membrane. In photosynthesis, energy absorbed from light is used to pump protons across the membrane. In both cases the protons are pumped from the inside, cytoplasmic face, to the outside, exoplasmic face, of the membrane. In addition to the production of ATP, this proton motive force can also be used by the cell to drive other processes such as the rotation of flagella to allow bacteria to swim around in solution or to drive the transport of species across the cell membrane against the existing concentration gradient. Clearly, an essential feature of this energy coupling between the transduction and its use in ATP synthesis or other processes is the presence of an insulating, closed membrane which is impermeable to the transport of protons, since without this it would not be possible to build up a transmembrane proton motive force. 1.3.1 The Proton Motive Force In this section we consider the thermodynamics of the proton motive force. The proton motive force is a combination of the potential difference across the membrane and the difference in proton concentration across the membrane. Both contribute to the available free energy. Consider an impermeable membrane separating two solutions, a and b, as shown in Figure 1.3. The electroHþ ðaÞ, is chemical potential of the proton in solution a, m given by  H þ ðaÞ ¼ m0H þ ðaÞ þ RT ln aH þ ðaÞ þ FfðaÞ m

ð1:1Þ

where m0Hþ ðaÞ is the chemical potential of the proton in solution a under standard conditions, R is the gas constant, T the temperature in Kelvin, aHþ ðaÞ is the activity of protons in solution a, F the Faraday constant and fðaÞ the potential. Notice that there are three contributions to the electrochemical potential of the proton in solution: a chemical term given by m0Hþ ðaÞ, an activity (or concentration) dependent term, and a term which depends on the potential. Similarly, for the protons in solution b, we can write  H þ ðbÞ ¼ m0H þ ðbÞ þ RT ln aH þ ðbÞ þ FfðbÞ m

ð1:2Þ

4

Bioelectrochemistry

Figure 1.2 A general overview of: (A) mitochondrial respiration and (B) photosynthesis.

At equilibrium, by definition the electrochemical potentials in the two solutions will be equal. That is  H þ ðbÞ  H þ ðaÞ ¼ m m

ð1:3Þ

However, a living cell is not at equilibrium. The difference in the electrochemical potential of the proton across the membrane, D mHþ , is a measure of the distance of the system from equilibrium and is given by H þ ¼ m  H þ ðaÞ  m  H þ ðbÞ Dm

We can simplify Equation (1.5) since the standard chemical potential of the proton is the same in both solutions, thus

ð1:4Þ

m0H þ ðaÞ ¼ m0H þ ðbÞ

After collecting together terms Equation (1.5) becomes  H þ ¼ RT ln Dm



 aH þ ðaÞ þ FDf aH þ ðbÞ

ð1:8Þ

The proton motive force itself is then

 H þ ¼ m0H þ ðaÞ þ RT ln aH þ ðaÞ þ FfðaÞ Dm  m0H þ ðbÞ  RTln aH þ ðbÞ  FfðbÞ

ð1:7Þ

or  H þ ¼  2:303RTDpH þ FDf Dm

so that

ð1:6Þ

ð1:5Þ

pmf ¼

H þ Dm  2:303RTDpH ¼ þ Df F F

ð1:9Þ

Bioenergetics and Biological Electron Transport

NH2

∆φ N O

(α)

5

N

N

O

N

O P O P O CH2 O H H O O H H OH OH Adenosine Diphosphate (ADP)

(β)

NH2 N O

H

+

H

O

N

N

O P O P O P O CH2 O H H O O O H H OH OH Adenosine Triphosphate (ATP)

+

Figure 1.3 Scheme for a transmembrane potential.

At room temperature pmf=mV ¼  59  DpH þ Df

O

N

ð1:10Þ

Thus the proton motive force is made up of two components: the contribution from the difference in proton concentration across the membrane, and the contribution from the potential difference across the membrane. Thus if the membrane is permeable to Cl or if Hþ exchanges with another cation (such as Kþ) the contribution from the potential difference Df will be small but DpH can still be large. This is the situation for the thylakoid membrane in photosynthesis. In contrast, if the membrane is impermeable to anions, Df can make a more significant contribution. This is the case in the respiring mitochondrion where the total proton motive force of around 220 mV is made up of a transmembrane potential of 160 mV (with the inside of the mitochondrion at a negative potential with respect to the outside – the protons are pumped from inside to out) together with a 60 mV contribution from the one unit pH difference across the membrane. 1.3.2 The Synthesis of ATP The second part of the story is the synthesis of ATP from ADP and inorganic phosphate and we now turn our attention to the thermodynamics of this process. Adenosine triphosphate, ATP (Figure 1.4), is found in all types of living organism and is the universal mode of transfer-

Figure 1.4 The structures of adenosine diphosphate (ADP) and adenosine triphosphate (ATP).

ring energy around the cell in order to drive all the endergonic (DGrxn > 0) reactions necessary for life. These include the synthesis of cellular macromolecules, such as DNA, RNA, proteins and polysaccharides, the synthesis of cellular constituents, such as phospholipids and metabolites, and cellular motion including muscle contraction. In humans it is estimated that on average 40 kg of ATP are used every day corresponding to 1000 turnovers between ADP, ATP and back to ADP for every molecule of ADP in the body each day [5]. In ATP the energy is stored in highenergy phosphoanhydride bonds and hydrolysis of these bonds to produce ADP or AMP (adenosine monophosphate) releases this energy ATP4  þ H2 O ! ADP3  þ H þ þ HPO4 2 

ð1:11Þ

ATP4  þ H2 O ! AMP2  þ H þ þ HP2 O7 3 

ð1:12Þ

or

where HPO42 is inorganic phosphate and HP2O73 is inorganic pyrophosphate. In the case of Reaction (1.12), the inorganic pyrophosphate produced is hydrolysed to inorganic phosphate by the enzyme pyrophosphatase. Both reactions, (1.11) and (1.12), have a free energy change of 30.5 kJ mol1 in the standard state at pH 7. When we take account of the actual concentrations of the different species (2.5 mM for ATP, 0.25 mM for ADP and

6

Bioelectrochemistry

2.0 mM for HPO42) this gives a value of about 52 kJ mol1 in the living cell [6]. If we assume a transmembrane proton motive force of, say, 200 mV this corresponds to a free energy change for each proton translocated across the membrane of 19.3 kJ mol1. Thus it is necessary to transfer at least two protons across the membrane for each ATP molecule synthesised.

H OH

Glucose

HO HO

HO OH OH H H +2NAD + +2ADP 3- +2HPO 42H

multiple steps and intermediates

1.4 ELECTRON TRANSPORT CHAINS We now turn our attention in more detail onto the electron transport chains in mitochondria, in the chloroplast and in bacteria and focus on the processes occurring in these electron transport chains. As we have seen, electron transport in these systems is central to the process of energy generation in living systems. We can therefore expect these systems to have evolved to operate efficiently and it is of interest to study the way that they operate and the underlying principles involved. We begin by considering the mitochondrial electron transport chain. 1.4.1 The Mitochondrion The main source of energy in non-photosynthetic cells is glucose. The standard free energy for the complete oxidation of glucose to carbon dioxide and water is 2870 kJ mol1 and in the cell this is coupled to the synthesis of up to 32 moles of ATP for every mole of glucose oxidised. In eukaryotic cells, the first stages of this process occur in the cytosol, the solution contained by the outer cell membrane, where two moles of ATP are generated by the conversion of glucose to two molecules of pyruvate in a process called glycolysis (Figure 1.5). In addition to two molecules of ATP, the process generates two molecules of NADH. The pyruvate generated in the cytosol is transported to the mitochondria where up to 30 further ATP molecules are generated by the complete oxidation of the pyruvate to carbon dioxide. In addition, the two molecules of NADH formed in glycolysis reduce two molecules of NADþ within the mitochondrion, which are then oxidised back to NADþ by oxygen as part of the mitochondrial electron transport chain. Thus the mitochondrion is the central power plant, or more precisely the central fuel cell, which powers the eukaryotic cell, and these cells generally contain hundreds of mitochondria [7]. In humans, for example, it has been calculated that at rest the typical transmembrane current, summed over all the mitochondria, amounts to just over 500 A (assuming a power consumption of 116 W and a transmembrane potential of 0.2 V) [8]. Given the central role of mitochondria in energy production, it is not surprising that mitochondrial

O

Pyruvate (2 molecules) O

O

O

O

O +2NADH +2ATP4-

Figure 1.5 The overall reaction for glycolysis.

defects are implicated with a wide range of degenerative diseases [7]. It is also worth noting that mitochondria are also essential in the photosynthetic cells of plants for the production of ATP during dark periods and for the generation of ATP in all non-photosynthetic plant cells (such as root cells). The mitochondrion is around 1 or 2 microns in length and 0.1 to 0.5 microns in diameter and is therefore one of the larger organelles in the eukaryotic cell. The mitochondrion, Figure 1.6, contains two separate membranes [9]. The outer membrane is made up of about 50 % lipid and 50 % of proteins called porins, which allow molecules with molecular weight up to 10 000 Da to pass through. The inner membrane is much less porous and is about 20 % lipid and 80 % protein. The inner membrane has a large number of invaginations, called cristae, which increase the surface area of the membrane and it is across this inner membrane that electron and proton transport occur in the mitochondrial electron transport chain. The

Figure 1.6 The structure of the mitochondrion.

Bioenergetics and Biological Electron Transport

O

HCoSA

HPO4 2+ GDP

CO 2

GTP CO2 + HSCoA

SCoA O

OH

NAD +

NADH O

7

3NAD+

3NADH

FAD

FADH 2

Figure 1.7 The citric acid cycle. HCoSA is free coenzyme A; GDP and GTP are guanosine diphosphate and guanosine triphosphate, respectively.

solution inside the mitochondrion has a high protein content, around 50 % by weight, and therefore is quite viscous. Broadly, three processes occur within the mitochondria. First, the oxidation of pyruvate to carbon dioxide with the generation of NADH and FADH2 through the citric acid cycle (Figure 1.7). Second, the oxidation of NADH and FADH2, by molecular oxygen in the mitochondrial electron transport chain, to generate a proton motive force across the inner mitochondrial membrane. Third, the generation of ATP from ADP by F0F1 ATPase, driven by the proton motive force across the inner membrane. This ATP is then exported from the mitochondrion to drive processes in other parts of the eukaryotic cell. In this chapter we will concentrate on the last two of these processes since they are essentially electrochemical, rather than chemical, in their nature. Here we examine the mitochondrial electron transport chain, a description of the F0F1 ATPase is given in a later section. Overall, the oxidation of one molecule of glucose produces 10 molecules of NADH and two molecules of FADH2. This process is reasonably efficient and incurs only about a 10 % energy loss from that originally available in the glucose. In the electron transport chain, the reduced coenzymes are reoxidised in several distinct steps by molecular oxygen rather than in a single step – by using the electron transfer chain, the energy is released in a number of small, and therefore thermodynamically more efficient, steps. Overall, the oxidations of NADH and FADH2 are strongly exergonic (DGrxn < 0) processes NADH þ H þ þ 1 =2 O2 ! NAD þ þ H2 O DG ¼  220 kJ mol  1

ð1:13Þ

FADH2 þ 1 =2 O2 ! FAD þ H2 O DG ¼  182 kJ mol  1 ð1:14Þ Since the conversion of ADP to ATP requires 30.5 kJ mol1, the oxidation of one molecule of NADH or FADH2 is sufficiently exergonic to generate more that one molecule of ATP.

Figure 1.8 An overview of the mitochondrial electron transport chain: (a) Electron transfer from NADH to molecular oxygen; (b) electron transfer from FADH2 and succinate to molecular oxygen. I, II, III and IV are the four major transmembrane protein complexes of the mitochondrial electron transport chain. The processes leading to the pumping of protons across the membrane are indicated by the bold arrows. The broken arrow shows the pathway of sequential electron transfer down the chain. CoQ and Cyt c are coenzyme Q and cytochrome c respectively.

Figure 1.8 gives an overview of the mitochondrial electron transport chain [2,10,11]. It comprises several large proteins which span the inner mitochondrial membrane. As electrons are passed along the chain from NADH at one end, to oxygen at the other, protons are pumped across the membrane at several points as shown

8

Bioelectrochemistry

in the figure. Oxidation of NADH occurs at the NADHCoQ reductase complex (complex I in Figure 1.8a). This process is accompanied by the transfer of four protons across the membrane and the electrons from the NADH are passed to a molecule of coenzyme Q, CoQ, a hydrophobic quinone, which takes the electrons from the NADH-CoQ reductase complex and passes them to the CoQH2–cyt c reductase complex (complex III in Figure 1.8). Note that the structures of the individual redox centres in the different parts of electron transport chains are discussed in detail later in this chapter – for now we concentrate on the larger functional picture rather than the molecular detail. The reduced coenzyme Q passes the electrons to the CoQH2–cyt c reductase complex, where a further pair of protons are pumped across the membrane and the electrons are passed to two molecules of cytochrome c, cyt c, a soluble electron transfer protein. The cytochrome c passes the electrons to the cyt c oxidase complex (complex IV in Figure 1.8) where a further two protons are pumped across the membrane and the electrons end up on oxygen, producing water. The electrons from FADH2 are fed into the electron transfer chain in a similar way (Figure 1.8b). FADH2 is oxidised to FAD by the succinate–CoQ reductase complex (complex II in Figure 1.8b), with the generation of one molecule of reduced coenzyme Q. No protons are pumped across the membrane by this reaction. The reduced coenzyme Q produced from FADH2 joins that from NADH in passing its electrons to the CoQH2–cyt c reductase complex and thence to the cyt c oxidase complex and ultimately to molecular oxygen (Figure 1.8b). The succinate-CoQ reductase complex (complex II in Figure 1.8b) also catalyses the oxidation of succinate, produced by the

citric acid cycle within the mitochondrion, to fumarate, with the generation of one molecule of reduced coenzyme Q which participates in the electron transport chain. The four protein complexes, I to IV in Figure 1.8, are large multiunit proteins each containing several redox prosthetic groups (Table 1.1) and within the individual protein complexes, the redox prosthetic groups are carefully arranged in three dimensions so that the electrons are passed in an ordered fashion from one redox component to the next. Much progress has been made in the last few years in determining the structures of these large, membrane-bound proteins and the details of their operation. In the following sections we discuss electron transfer in each of the complexes in turn.

1.4.2 The NADH–CoQ Reductase Complex The NADH–CoQ reductase complex, or complex I, is found in the mitochondria of eukaryotes and in the plasma membranes of purple photosynthetic bacteria and respiratory bacteria. The complex (Figure 1.9a) has an L-shape with two major sub-units, one predominantly within the membrane and the other protruding into the inner mitochondrial space containing the NADH reaction site [12–15]. The NADH–CoQ reductase complex is the most complex and largest of the proton pumping enzymes in the mitochondrion and is made up of about 30 separate sub-units; it is also, because of this complexity, the least well understood. The NADH reacts with a flavin mononucleotide (FMN) prosthetic group. From here the electrons are passed to several (eight or nine) iron–sulfur centres.

Table 1.1 The four protein complexes of the mitochondrial electron transport chain. Complex I

II

III

IV

NADH–CoQ reductase

Succinate–CoQ reductase

CoQH2–cyt c reductase

Cyt c oxidase

RMM/kDa

Redox centres

>900

FMN

120

8 or 9 Fe-S 2 quinones FAD

240

3 Fe-S 1 Heme b 2 Heme b

204

1 Heme c 1 Fe-S 2 Heme a 2 Cu

Comments

Ref þ

transmembrane, pumps H , 43–46 subunits

[12–15]

membrane bound, does not pump Hþ, 4 subunits

[11,22–24]

transmembrane, pumps Hþ, 11 subunits, exits in membrane as functional dimmer

[28,31]

transmembrane, pumps Hþ, 13 subunits

[37–39]

Bioenergetics and Biological Electron Transport

9

Here the subscripts ‘inside’ and ‘outside’ refer to the location of the proton with respect to the inner mitochondrial membrane. At present there is not a crystal structure for the NADH–CoQ reductase complex and the precise pathway of electron transfer within the complex and the mechanism and stoichiometry of proton transport are currently not well established, although a crystal structure of the hydrophilic domain of complex I from Thermus thermoplilus has recently been obtained [17]. For further details of the NADH–CoQ reductase complex, readers are directed to several recent reviews [13,18–21].

1.4.3 The Succinate–CoQ Reductase Complex

Figure 1.9 (a) Schematic of the NADH–CoQ reductase complex (complex I) showing the organisation of the redox sites. N-1, N-2, N-3 and N-4 are iron sulfur clusters, Qi is an internal quinone, FMN is flavin mononucleotide and CoQ is coenzyme Q. The full arrows show the direction of electron transfer and the broken arrows the direction of proton transfer (based on Hofhaus et al. [14]). (b) The arrangement of the electron transfer chain in the hydrophilic (peripheral) arm of the complex (based on Hinchcliffe and Sazanov [16]. N1a and N1b are two iron two sulfur clusters; N2, N3, N4, N5, N6a/b, N6b/a and N7 are four iron four sulfur clusters. The suggested electron transport pathway is shown by the arrows.

A recent study by Hinchliffe and Sazanov [16] has found a chain of eight iron–sulphur clusters in the hydrophilic domain of the complex with edge-to-edge spacings of less than 1.4 nm with the ninth iron–sulfur cluster somewhat further away (Figure 1.9b). The eight iron–sulfur clusters form an electron transport chain 8.4 nm long, which is believed to connect the two catalytic sites of the enzyme. Reduction of coenzyme Q occurs at the part of the protein complex which is within the membrane. Overall the reaction of one molecule of NADH generates one molecule of reduced coenzyme Q and, it is suggested, pumps four protons across the membrane þ þ NADH þ CoQ þ 5Hinside ! NAD þ þ CoQH2 þ 4Houtside

ð1:15Þ

The succinate–CoQ reductase complex, also referred to as complex II, spans the mitochondrial membrane, but does not pump protons across it – the free energy released by the reaction of succinate with coenzyme Q is insufficient to drive the transfer of a proton across the membrane [11]. The complex comprises two hydrophilic sub-units and one or two hydrophobic sub-units, which are associated with the membrane [22–24]. Succinate reacts at a flavin, FAD, prosthetic group in one of the hydrophilic sub-units, of the complex located on the inside of the membrane. From here the electrons are passed one at a time to iron–sulfur centres and thus to the coenzyme Q reduction site (Figure 1.10). Succinate–CoQ reductase complexes from some species also contain cytochrome b560 redox sites [22,23,25].

1.4.4 The CoQH2–Cyt c Reductase Complex The CoQH2-cyt c reductase complex, also referred to as the cytochrome bc1 complex or complex III, is a homodimeric transmembrane protein complex that takes electrons from the reduced coenzyme Q, produced by the NADH–CoQ reductase and succinate–CoQ reductase complexes, and passes the electrons to cytochrome c, a water soluble 13 kDa electron transfer protein through the so-called Q cycle [26,27]. In doing so it pumps two protons across the membrane and releases another two from the reduced quinone þ CoQH2 þ 2cyt cðFe3 þ Þ þ 2Hinside þ ! CoQ þ 2cyt cðFe2 þ Þ þ 4Houtside

ð1:16Þ

Figure 1.11 summarises the scheme. In the complex there are three catalytic sub-units which contain two cytochrome b redox centres (b-type hemes), a cytochrome c1 redox centre and a two iron two sulfur, Fe2S2, centre, respectively.

10

Bioelectrochemistry

Figure 1.10 A schematic of the succinate–CoQ reductase complex (complex II) showing the organisation of the redox sites. The arrows show the direction of electron transfer. FeS are iron–sulfur clusters, FAD is flavin dinucleotide, cyt b560 is cytochrome b560 and CoQ is coenzyme Q (based on H€agerh€all and Hederstedt [23]).

The reduced coenzyme Q, CoQH2, is oxidised in two steps with one electron being transferred to a high potential redox chain to give the semiquinone, CoQ , which then gives up a second electron to a separate low-potential redox chain in what appears to be a concerted electron transfer, since no intermediate semiquinone can be detected. In the first electron transfer the electron is transferred to the Fe2S2 cluster and then to the cytochrome c1 from where it is transferred to the soluble cytochrome c electron transfer protein. This is referred to as the high-potential redox pathway, because the redox potentials of the Fe2S2 cluster and the cytochrome c1 redox centres are significantly more positive (300 mV vs NHE) than those of the cytochrome b centres in the other pathway (50 and 70 mV vs NHE). The second electron is transferred from the semiquinone, CoQ , to one cytochrome b and then on to a second cytochrome b, at a site within the protein complex on the other side of the membrane, where coenzyme Q is reduced in two steps to CoQH2, the fully reduced form. This makes up the Q cycle and leads to pumping of protons across the membrane. Crystallographic studies by Zhang et al. [28] suggest that a significant conformational change in the protein is associated with this direction of the electrons from the reduced coenzyme Q [27]. In one conformation the Fe2S2 cluster is close enough to the coenzyme Q binding site to pick up an electron. In the second conformation the Fe2S2 cluster swings away from the coenzyme Q binding site, moving through about 1.6 nm, and approaches close enough to the cytochrome c2 heme to allow electron transfer. At the same time the Fe2S2 cluster is too far from the coenzyme Q binding site to collect the second electron, which is therefore passed to the heme of the cytochrome b. For a discussion of possible short circuits between these two pathways in the Q cycle see Osyczka et al. [29,30]. The crystal structures of CoQH2–cyt c reductase complexes from several organisms have been solved at high resolution [28,31], so that we know quite a lot about the relative organisation and distances between the redox groups within the protein. 1.4.5 The Cyt c Oxidase Complex

Figure 1.11 A schematic of the CoQH2–cyt c reductase complex (complex III) showing the organisation of the redox sites. The arrows show the direction of electron and proton transfer. Qo and Qi represent the coenzyme Q, CoQ, reaction sites on the outside and inside of the membrane respectively, cyt c1, cyt bL and cyt bH are three cytochrome centres with bL and bH being the low and high potential cytochrome b centres respectively, Fe2S2 is a two iron, two sulfur cluster. The overall complex is a homodimer, the electron and proton transfer chains are only shown for one half for clarity.

The cyt c oxidase complex, or complex IV, is the terminus for the mitochondrial electron transport and the site for the reduction of molecular oxygen to water [32–36]. The mitochondrial complex has 13 sub-units, two of which have catalytic functions, while the others are involved in the binding of the active sub-units and a lipid molecule [37–39]. The first of the catalytic sub-units contains an unusual Cu centre with two Cu atoms, which is thought to be the reaction site for the cytochrome c. The second of the catalytic

Bioenergetics and Biological Electron Transport

Figure 1.12 A schematic of the cyt c oxidase complex (complex IV) showing the organisation of the redox sites. Hemes a and a3 are the two heme centres associated with the copper b site, Cub2þ; Cua2þ are the two copper a centres (based on Faxen et al. [40]).

sub-units contains two type-a heme centres; one of these accepts electrons from the first sub-unit. The second type-a heme forms part of a binuclear centre with a Cu centre. This binuclear Cu/heme-a centre is the site for oxygen reduction (Figure 1.12). Transfer of electrons from the reduced cytochrome c, produced by the CoQH2–cyt c reductase complex, through the Cyt c oxidase complex to molecular oxygen, leads to the transfer of a further two protons across the inner mitochondrial membrane for every pair of electrons transferred þ 2 cyt cðFe2 þ Þ þ 4Hinside þ 1 =2 O2 þ ! 2 cyt cðFe3 þ Þ þ H2 O þ 2Houtside

ð1:17Þ

The precise molecular mechanism by which the protons are pumped across the membrane has been investigated by Faxen et al. [40]. From an electrochemical perspective, the detailed mechanism of the reduction of molecular oxygen to water by the cyt c oxidase complex is of particular interest, since the design of efficient catalysts for four electron reduction of oxygen at neutral pH remains a very significant impediment to the development of PEM fuel cells. At present, although crystal structures have been obtained for cyt c oxidase complexes, the resolution is not sufficient to

11

Figure 1.13 The mechanism for oxygen reduction in the cyt c oxidase complex proposed by Michel et al. [35]. O is the oxidised form, E the one-electron reduced form, R the two-electron reduced form, and A the product formed on oxygen binding. PM and PR are peroxy intermediates, alternative structures are given in the box on the right assuming that the OO bond is already broken in these states; the missing electron could be provided by a porphyrin ring (Por þ), an amino acid residue (res þ), a copper b site (Cu3þ), or the heme a3-Fe (Fe5þ). F is the oxyferryl state and H the hydroxyl state formed after protonation of the iron-bound oxygen.

fully define the geometry of the oxygen binding site. Furthermore the X-ray structures show only a snapshot of the structure and cannot reveal the dynamics of protein movement during the catalytic cycle. Two mechanisms have been proposed to describe oxygen reduction by the binuclear Cu/heme-a site. As shown in Figure 1.13, Michel et al. [35] suggest that the oxygen binds to the heme iron (A) and then forms a peroxy intermediate (PM), by transfer of an electron from the Cu, followed by the addition of a second electron and two protons to produce an oxoferryl state (F). Further addition of a proton and electron gives a hydroxy state (H). Protonation and two further electron transfers return the system to the starting, doubly reduced state (R). Michel et al. also suggest that there could be an alternative route if one assumes that the O¼O bond is broken at an earlier stage. In this case there are several possibilities for the intermediate species, as shown in the box in Figure 1.13, depending on where the additional electron is taken from. In contrast, Wikstr€ om [41] has proposed a different mechanism in which the two peroxy intermediates, PR and PM, correspond to ferryl, Fe(IV), species where the additional electron is provided by a nearby tyrosine (YOH) (Figure 1.14). In both cases the close proximity of the heme iron and the Cu,

12

Bioelectrochemistry

Figure 1.15 The respiratory electron transport chain of marine Vibrio based on the work of Unemoto [45]. Q is ubiquinone, FNM is flavin mononucleotide and FAD is flavin dinucleotide.

Figure 1.14 The mechanism for oxygen reduction in the cyt c oxidase complex proposed by Wikstr€om [41]. The boxes represent the binuclear centre with the nearby tyrosine residue (YOH). O is the oxidised form, R the two-electron reduced form, A the product formed on oxygen binding and PM and PR are peroxy intermediates.

they are within 0.52 nm of each other, is important for the catalysis. 1.4.6 Electron Transport Chains in Bacteria In comparison to mitochondria, bacteria tend to contain redundant electron transport systems [42]. As a result, bacteria can grow under a variety of conditions and can switch between different branches of their electron transport chains depending upon the conditions. Bacteria utilize a range of electron donors in energy generation, together with either oxygen as the ultimate electron acceptor, in aerobic respiration, or other species such as NO3 or SO42, in anaerobic respiration, although in these cases less energy is generated. The electron transport chains in bacteria are very similar to those in the mitochondrion and the principles are the same [43], although the polypeptide composition of the electron transport proteins in bacteria is usually simpler than those in mitochondria and the proteins involved are located in the cell membrane itself (prokaryotes do not contain separate mitochondria). Sequential electron transfers between different components in the electron transport chain leads to the pumping of protons across the membrane. This establishes a proton motive force which the cell uses to synthesise ATP from ADP. In addition, the different protein complexes and prosthetic redox centres are closely related to those of the mitochondrion, although

there are differences in the detailed structures. In fact, this is not surprising since it is now accepted that the mitochondria found in eukaryotes arose more than 1 billion years ago when an energetically inefficient eukaryotic cell was invaded by a more energy efficient bacterium, a process called endosymbiosis. Subsequent transfer of much of this bacterial genetic information to the nucleus of the eukaryotic cell led to the invading symbiotic bacterium being transformed into the structure we know today as the mitochondrion [44]. As a consequence, mitochondrial DNA (mtDNA) which codes for the four respiratory complexes (I to IV) and the ATP synthase found in the mitochondrial membrane, is distinct from the rest of the DNA of the organism, uses a different DNA code, and is strictly maternally inherited [7,44]. As an illustration, Figure 1.15 shows the respiratory chain of marine Vibrio [45]. Notice, in this case, that there are three main complexes, an NADH–quinone reductase, a quinol oxidase, and an NADH–quinone reductase. These are linked together by quinones. In this case, the organism pumps both Hþ and Naþ across the membrane to generate both a proton motive force and a transmembrane difference in the electrochemical potential of Naþ. Escherichia coli is able to assemble specific respiratory chains by the synthesis of the necessary dehydrogenase and reductase enzymes in response to the conditions in which it finds itself. As an example, Figure 1.16 shows an electron transport chain for Escherichia coli during anaerobic respiration [46,47]. In this case, nitrate replaces oxygen as the ultimate electron acceptor. 1.4.7 Electron Transfer in Photosynthesis A very similar situation pertains in photosynthesis as in the mitochondrial electron transport chain, except that this time the energy is provided by light rather than by

Bioenergetics and Biological Electron Transport

Figure 1.16 The electron transport chain for Escherichia coli during anaerobic respiration (based on Smith and Wood [70], Dym et al. [47] and Bertero et al. [46]). FAD is flavin, FeS are iron–sulfur clusters, Q is menaquinone, cyt b is a b-type cytochrome and Mo is molybdopterin-guanosine-dinucleotide.

glucose. However, the molecular organisation and the structures of many of the components involved are very similar, and the governing principle, that sequential electron transfers lead to the generation of a proton motive force across the energy transducing membrane by pumping protons across the membrane, remains the same. In this section we will look at the photosynthetic electron transport chain found both in plants and bacteria. Again, as in respiration, the proton motive force set up by the photosynthetic membrane is used to drive the formation of ATP from ADP. In plants, photosynthesis occurs in the chloroplasts. These are large organelles found mainly in the leaf cells of the plant. The principle products of the photosynthetic metabolism of carbon dioxide are C6 sugars in the form of sucrose and starch. The sucrose is water soluble and is transported from the chloroplast to other parts of the plant to provide energy for metabolism. The starch is stored within the leaf. The chloroplast (Figure 1.17) has three membranes. The outermost membrane of the chloroplast, as with the mitochondrion, contains a large number of porin proteins, which make it readily permeable to low molecular weight species, such as sucrose. Inside this, the second membrane is the primary permeability barrier of the chloroplast and contains various permeases, proteins which control ingress and egress of species from the chloroplast interior. Unlike the mitochondrion, the energy transducing membrane of the chloroplast, called the thylakoid membrane, is separate from the inner membrane of the organelle. The thylakoid membrane is where the chlorophyll is located and is the site of energy conversion. Again, note that as for the mitochondrion, the thylakoid membrane totally encloses a volume of solution, called the lumen, within

13

Figure 1.17 The structure of the chloroplast.

the chloroplast, so that a proton motive force can be generated across the membrane during photosynthesis. Within the chloroplast these thylakoid membranes frequently form flattened, pancake-like structures, called grana, which then form into stacks. The thylakoid membrane in algae and higher plants contains two photosystems referred to as photosystem I and photosystem II or PS I and PS II. Both photosystems contain chlorophyll and under irradiation, the absorption of light within the photosystem leads to charge separation across the thylakoid membrane, with positive charge being drive to the lumen side of the membrane and negative charge to the stroma side of the membrane. The stroma side is the side of the thylakoid membrane which is on the outside – that is the solution contained within the chloroplast. Photochemically driven charge separation in photosystems I and II is coupled together by a quinone cycle and a cytochrome bf complex (Figure 1.18). Overall the photochemically driven reaction is hv

þ þ 2H2 O þ 2NADP þ þ 2Houtside ! O2 þ 2NADPH þ 4Hinside

ð1:18Þ where electrons are transferred from water to NADPþ. This reaction is essentially the reverse of the reaction of respiration in the mitochondrion (NADPH is closely related in structure to NADH, see below). The measured values for the quantum requirement (that is the number of photons required for each molecule of oxygen produced) in intact leaves, under ideal conditions, are typically nine or ten, close to the theoretical value of eight with one photon absorbed by each of the two photosystems for each

14

Bioelectrochemistry

Figure 1.18 The overall scheme for the photosynthetic electron transfer pathway showing the two photosystems, photosystem I and photosystem II, and the direction of electron and proton transfer. NADPþ is nicotinamide adenine dinucleotide phosphate and chl is chlorophyll.

The overall efficiency for this process, calculated in terms of the energy stored over the energy absorbed in the form of light, comes out at about 27 % [6]. In this section we will focus specifically on the electron transfer reactions which occur within the thylakoid membrane; those who would like to know more about the photochemical, as distinct from the electrochemical, processes which accompany photosynthesis are directed to an excellent recent text by Blankenship [6]. Figure 1.19 shows the different components in the electron transport chain of the thylakoid electron transport chain [2,48–50]. The process of photosynthesis begins with the absorption of light by the light harvesting complex of photosys6CO2 þ 18ATP4  þ 12NADPH þ 12H2 O tem II. This energy is passed to the photoreaction centre hv ! C6 H12 O6 þ 18ADP3  þ 12NADP þ þ 18HPO4 2  þ 6H þ of photosystem II where it drives charge separation across ð1:19Þ the thylakoid membrane leading to the oxidation of water electron transported along the chain [6]. This photochemical reaction generates a proton motive force across the thylakoid membrane in which the inside, lumen, is positive. This proton motive force is used by the CF0CF1 complex, a large transmembrane protein complex embedded in the thylakoid membrane, to generate ATP from ADP in much the same way that the FoF1 ATPase complex in the mitochondrion generates ATP (see below). The ATP and NADPH generated by the photochemically driven reactions are used within the chloroplast to fix carbon dioxide and produce sugars

Figure 1.19 The different components of the photosynthetic electron transfer pathway, showing the light harvesting complexes, LHC, and photosystems, PS. The arrows show the direction of electron and proton transfer. PQ is plastoquinone, PC is plastocyanin, Fd is flavodoxin, PS I and PS II are photosystem I and II respectively, NFR is ferredoxin-NADP oxidoreductase, P600 and P700 are the reaction centre chlorophylls, Mn is the four manganese oxygen evolving complex, cyt f and cyt b are cytochromes, FeS is an iron–sulfur cluster and FAD is flavin adenine dinucleotide.

Bioenergetics and Biological Electron Transport

to molecular oxygen, the generation of a reduced molecule of ubiquinone (the equivalent of coenzyme Q in the plant system) and the translocation of protons across the membrane. The reduced ubiquinone (QH2) passes the electrons to the cytochrome bf complex and hence to plastocyanin, a water-soluble electron transfer protein with a single Cuþ/2þ redox site. At the same time protons are pumped across the membrane. From plastocyanin the electrons are passed to photosystem I, where further absorption of light drives electron transfer across the membrane and the reduction of NADPþ to NADPH. Photosystems I and II both have light-harvesting complexes associated with them, but the two are structurally different and are located in different parts of the thylakoid membrane. Photosystem II and its associated light-harvesting complex is located mainly in the stacked grana membranes, whereas photosystem I and the CF0CF1 complex are located in the parts of the membrane, the stroma, which link together the grana. The cytochrome bf complex is found in both grana and stroma membranes [51]. Since the photosystems are located in physically separate parts of the thylakoid membrane, relatively long range (tens of nm) diffusive transport of electrons by the plastoquinone associated with the membrane and the plastocyanin in the lumen play an important part in the overall process. Comparing the overall picture shown in Figure 1.19 for photosynthesis with that for the mitochondrial electron transport chain, Figure 1.8, reveals some significant similarities. Thus, quinone redox species, either coenzyme Q or plastoquinone, play an important role in proton transfer across the membrane; water soluble one-electron transfer proteins, either cytochrome c or plastocyanin, are used to couple together electron transfer between large transmembrane proteins, and, as we shall see below when we consider the components of the photosynthetic electron transport chain in more detail, there are striking similarities between the CoQH2–cyt c reductase complex in mitochondria and the cytochrome bf complex in photosynthesis. These significant similarities in the operation of the two major energy transducing systems in biology strongly suggest a common evolutionary origin. 1.4.8 Photosystem II Photosystem II is a multisub-unit protein complex which is found embedded in the thylakoid membranes of higher plants as well as algae and cyanobacteria [52,53]. Associated with photosystem II there is a light-harvesting complex, which contains an array of chlorophyll a, chlorophyll b and carotenoid pigments (Figure 1.20) [54–56]. The role of the light-harvesting complex is to

15

capture the energy from incoming light and funnel it to the photoreaction centre where charge separation occurs, it also has a role in the non-radiative dissipation of excess excitation energy to protect the system from damage at high light levels. The light-harvesting complex is necessary because, even in full sunlight, the light falling on the plant leaf represents a fairly dilute energy source [6]. The different pigments in the light-harvesting complex pigments have the effect of extending the range of wavelengths of light that the plant can absorb. When light is absorbed by the pigment array of the light-harvesting complex, the energy is passed by rapid resonant energy transfer in less than 1 ns to a pair of the chlorophyll a pigments in the photoreaction centre of photosystem II. This special pair of chlorophyll a molecules, P680, play a central role in energy transduction in photosystem II, because it is at this stage that charge separation occurs (Figure 1.21). The Photosystem II photoreaction centre contains these two special chlorophyll a molecules (P680) together with two other chlorophylls, two pheophytin molecules (pheophytin is a metal-free chlorophyll, where the Mg2þ is replaced by two protons) and two quinones, all arranged to form an efficient electron transport chain [55,57–61]. Absorption of a photon with wavelength below 680 nm (corresponding to an energy of 176 kJ mol1) generates the oxidised form of the chlorophyll a, P680þ, by electron transfer via pheophytin and a quinone to the terminal quinone acceptor molecule on the outer surface of the thylakoid membrane. The P680þ is reduced by electron transfer from the oxygen evolving complex of photosystem II located on the inner side on the thylakoid membrane. This oxygen evolving complex contains a cluster of four manganese ions as well as bound chloride and calcium ions. The oxidation of water to molecular oxygen is a four-electron process 2H2 O þ 4e ! O2 þ 4H þ

ð1:20Þ

The cluster of four manganese ions in the oxygen evolving complex therefore cycles through four different oxidation states in order to couple the one-electron transfer to P680þ to the four-electron oxidation of water [62]. The precise details of the structure of the oxygen evolving complex remain the subject of debate [55,59]. With the oxidation of one molecule of water to oxygen four protons are released on the inside of the membrane. Thus, photosystem II takes in energy from the light absorbed by the light-harvesting complex and uses it to produce oxygen and pump protons across the thylakoid membrane contributing to the proton motive force across the membrane.

16

Bioelectrochemistry

N

N Mg

N

O

N

Chlorophyll a

O

H 3COOC O

O

N

N Mg

N

O

H3COOC

Chlorophyll b N

O

O

β-carotene Figure 1.20 Structures of the light-harvesting pigments.

1.4.9 Cytochrome bf Complex The cytochrome bf complex transfers electrons from photosystem II to photosystem I by catalysing the oxidation of reduced plastoquinone by plastocyanin, Pc

þ QH2 þ 2Pc þ ! Q þ 2Pc þ 2Houtside

ð1:21Þ

In some cases, this reaction is accompanied, as in the corresponding reaction of the mitochondrial CoQH2–cyt c reductase complex, by the pumping of two additional protons across the membrane through the operation of a Q cycle. For the cytochrome bf complex this is not always the case, however, under some circumstances it can switch to a mechanism in which the two additional protons are not pumped across the membrane [49].

Bioenergetics and Biological Electron Transport

17

Figure 1.21 A schematic of photosystem II showing the arrangement of the redox sites (based on Loll et al. [55]). QA and QB are plastoquinones, Chl is chlorophyll, Pheo is pheophytin, P680 are a special pair of chlorophyll molecules where charge separation occurs, Tyr are tyrosine residues, Mn4Ca is the four manganese cluster of the oxygen evolving complex, Fe2þ is a non-heme iron.

The cytochrome bf complex (Figure 1.22) is made up of four subunits and contains a 2Fe2S centre, two b-type hemes and a c-type heme [48]. The 2Fe2S centre is the site for oxidation of the reduced plastoquinone. The two b-type hemes span the hydrophobic core of the complex and are the basis of the Q cycle (as in the mitochondrial CoQH2–cyt c reductase complex, see above). The crystal structure of the

Figure 1.22 A schematic of the cytochrome bf complex showing the arrangement of the redox sites (based on Kurisu et al. [63]). The enzyme is a homodimer, the electron transfer pathway is shown in one half for clarity. PC is plastocyanin, PQ is plastoquinone and Fe2S2 is an iron–sulfur cluster.

cytochrome b6f complex from a cyanobacterium has recently been reported [63]. However, the full details of electron transfer in the cytochrome bf complex between the reduced plastoquinone and the plastocyanin are, as yet, not clear. 1.4.10 Photosystem I Much of our present understanding of photosystem I is based on the crystal structure solution for photosystem I from the cyanobacterium Synechococcus elongatus, by Jordan et al. [50,58,64,65]. The core of photosystem I is substantially larger than the corresponding core of photosystem II, but despite this there is still a single pair of chlorophyll a molecules at the heart of the charge separation process. Photosystem I contains significantly more chlorophyll than photosystem II, with about 90 chlorophyll molecules associated with the light-harvesting complex and six associated with the electron transport chain. The chlorophylls in the light-harvesting complex are arranged, along with about 20 carotenoids, in two layers (Figure 1.23). Once absorbed by the array of pigments in the light-harvesting complex, the excitation energy is rapidly passed by resonant energy transfer to a special pair of chlorophylls, P700, located in the core of the photoreaction centre, which are the start of the electron transfer chain. The absorption of this pair of chlorophylls is at 700 nm, slightly red shifted from that in photosystem

18

Bioelectrochemistry

Figure 1.23 A schematic of photosystem II and its light-harvesting complex showing the arrangements of pigments and redox centres and the direction of energy (dotted arrows) and electron (solid arrows) transfers (based on Jordan et al. [64] and K€uhlbrandt [65]). PC is plastocyanin, Fd is ferredoxin, Chl is chlorophyll, PhQ is phylloquinone, Fe4S4 are iron–sulfur clusters and P700 are a special pair of chlorophylls where charge separation occurs.

II. In photosystem I charge separation occurs by electron  transfer from the excited chlorophyll, P700 , through a chain of four accessory chlorophylls and two phylloquinones to three 4Fe4S clusters, the last of which is located on the outside of the thylakoid membrane. The chlorophylls and phylloquinones are arranged in two branches and there is still controversy over whether both branches or only one is involved in the electron transport [64]. The resulting oxidised chlorophyll, P700þ, is reduced by plastocyanin on the inside of the membrane. Unlike photosystem II, proton transfer across the membrane does not accompany the electron transfer. From the terminal 4Fe4S cluster, the electron is transferred to ferredoxin, a small (11 kDa) soluble one-electron redox protein containing a 2Fe2S cluster complexed to four cysteines. The reduced ferredoxin transfers electrons to NADP reductase, a peripheral protein bound on the inside of the thylakoid membrane near photosystem I [48]. Its role is to link the one electron transfers from the ferredoxin, Fd, to the two-electron reduction of NADPþ þ NADP þ þ Houtside þ 2Fd ! NADPH þ 2Fd þ

ð1:22Þ

To do this the ferredoxin reductase contains a single FAD centre with two different binding sites for ferredoxin and NADPþ [66]. Recent crystallographic studies of photosystem I from a higher plant have shown strong similarities in structure and in the positions of almost all of the chlorophylls with those in the cyanobacterium, despite their evolutionary divergence around 1 billion years ago [67].

1.4.11 Bacterial Photosynthesis There are five different major types of bacteria: cyanobacteria, purple bacteria, green sulfur bacteria, green nonsulfur bacteria and heliobacteria, that are capable of photosynthesis [6]. Of these only one, the cyanobacteria, produce oxygen. Photosynthesis in green and purple bacteria does not generate oxygen, because they only contain one photoreaction centre, rather than the two found in green plants, cyanobacteria and algae [2]. Figure 1.24 shows the electron transport chain in purple bacteria. In this case, the process is cyclic with each cycle pumping protons across the membrane. The basic components of the chain strongly resemble those of the chloroplast discussed above [68]. Again, this similarity is not surprising, but demonstrates a common evolutionary origin; as for the mitochondrion, it is now generally accepted that the chloroplast has an endosymbiotic origin, a view supported by genetic analysis [6]. In contrast to the chloroplast, in bacteria the various components of the photosynthetic apparatus are located in the bacteria’s lipid bilayer cytoplasmic membrane. In most cases this is surrounded by a second, more permeable, membrane and a tough outer cell wall to provide mechanical stability. The bacterial reaction centre captures light and uses this to drive charge separation across the bacterial membrane taking electrons from a reduced soluble cytochrome and passing on to a quinone. The resulting reduced quinone reacts with the transmembrane cytochrome bc1 complex through a Q cycle leading to the pumping of protons

Bioenergetics and Biological Electron Transport

19

Figure 1.24 The photosynthetic electron transport chain in purple bacteria (based on [2]). Q, QA, QB, Qi and Qo are quinones, Chl are a special pair of chlorophylls, Pheo is pheophytin, Cyt is a soluble cytochrome. The associated light-harvesting complex is not shown.

across the bacterial membrane and returning the electrons to the cytochrome. The proton motive force created by this photochemically driven redox cycle is used by the bacterial F0F1 complex to drive the synthesis of ATP. Electrons can also flow through the photosynthetic electron transfer pathway of purple bacteria by a linear (as opposed to cyclic) pathway. In this case, the electrons are ultimately transferred to NADþ, generating NADH within the cell and pumping protons across the membrane at the same time. In this case, the electron required to reduce the oxidised chlorophyll in the photoreaction centre comes from hydrogen sulfide (producing elemental sulfur) or from hydrogen gas. 1.5 REDOX COMPONENTS As we have seen in the previous sections, energy transduction in living organisms, either by photosynthesis or respiration, proceeds via a sequence of ordered electron transfer reactions which generate a proton motive force across an impermeable membrane. If is also clear from our discussion so far that there are significant similarities between the respiratory electron transport chain in prokaryotic bacteria and in the mitochondria of eukaryotic cells, and the photosynthetic electron transport chains found in green and purple bacteria and in the chloroplast. In particular, despite the apparent complexity of the different electron transport chains, nature uses a relatively limited palette of redox active centres: heme, quinones, flavins, iron–sulfur clusters, etc. In this section, we describe the structures and electrochemical reactions of these different centres.

1.5.1 Quinones Quinones are two-electron, two-proton redox centres (Figure 1.25) for which the intermediate semiquinone radical is accessible and often reasonably stable so that they can undergo sequential one-electron oxidation or reduction reactions. Quinones are thus hydrogen atom carriers and the quinones involved in the energy transducing electron transport chains couple electron transport between large, transmembrane protein complexes such as photosystem II and the cytochrome bf complex or the NADH–CoQ reductase complex and the CoQ–cyt c reductase complex, and transport protons across the energy transducing membrane. Because the redox reaction of the quinones involves both electrons and protons, the redox potential of the couple is pH dependent, shifting by 59 mV for each unit change in pH at 298 K for the overall twoelectron, two-proton reaction. To achieve this, all the electron transport chain quinones have a long isoprenoid chain (Figure 1.26) which makes them lipid soluble so that they can freely diffuse in the lipid membrane. In coenzyme Q (also called ubiquinone because of its ubiquity), this chain comprises between six and 10 isoprenoid units, depending on the particular organism; in humans the chain is 10 isoprenoid groups in length. The redox potential for coenzyme Q is þ0.100 V vs SHE at pH 7. The structure of the plastoquinones involved in photosynthesis in the chloroplast is very similar to that of coenzyme Q (Figure 1.26), with only slight changes in the substitution of the quinone ring. These changes lead to a change in the redox potential of the plastoquinones to þ0.08 V vs SHE. Again there is an

20

Bioelectrochemistry

O H3CO

CH3

H3CO

CH2CH C

isoprenoid chain with between six and nine groups to ensure lipid solubility of the molecule. Some bacteria contain napthoquinones, such as menaquinone (redox potential þ0.07 V vs SHE at pH 7 [69]) (Figure 1.26), as well as coenzyme Q in their respiratory chains [70]. The principle, however, remains the same.

CH3 H2 Quinone C H

O

10

e

1.5.2 Flavins

O H3CO

CH3

H3CO

CH3 H2 CH2CH=C C H Semiquinone 10

O

e + 2H + OH H3CO

CH3 CH3 H2 CH2CH=C C H Hydroquinone

H3CO

Flavin adenine dinucleotide, FAD, and flavin mononucleotide, FMN, are also two-electron, two-proton redox centres [71]. The one-electron oxidation or reduction semiflavin intermediate is accessible and reasonably stable (Figure 1.27). As for the quinones, the redox potentials for flavin couples varies with pH. The two molecules, FAD and FMN, differ in that FAD has an additional phosphate, ribose and adenosine unit attached to the ribitol phosphate chain attached to on the flavin ring (Figure 1.28). This peripheral change, away from, and not conjugated to, the redox active part of the molecule does not alter the redox potential of the couple (0.21 V vs SCE at pH 7). However, the potentials of the flavin redox centre in different proteins differs widely, a point we return to below.

10

OH

1.5.3 NAD(P)H

Figure 1.25 The redox reactions for quinone showing the structures of the quinone, semiquinone and hydroquinone.

b-Nicotinamide adenine dinucleotide, NADþ, and bnicotinamide adenine dinucleotide phosphate, NADPþ,

O CH 3 CH 3 O

menaquinone, n = 4–3

CH2CH=C CH2 H n

O H 3CO

CH 3

H 3CO

CH2CH=C CH2 H 10

CH 3 O

ubiquinone

O H 3C CH 3 H 3C O

CH2CH=C CH2 H n

plastoquinone, n = 6–9

Figure 1.26 The structures of three quinones which are important components of electron transport chains.

Bioenergetics and Biological Electron Transport

O H3C

N

H3C

N

are two-electron, one-proton redox couples for which the intermediate radical forms are not readily accessible (Figure 1.29). NADH and NADPH act as hydride carriers in the biological system and generally undergo oxidation by hydride transfer in a single step. This hydride transfer can occur either to or from the alpha or beta face of the molecule. The choice of the particular face is determined by the binding of the NAD(P) within the active site of the enzyme, and is different for different enzymes. NADþ and NADPþ differ in that NADPþ has an additional phosphate on the ribose ring of the adenosine. Again, this is sufficiently removed from, and not conjugated to, the redox centre within the molecule, so that the redox potentials of the two couples are the same. The change does, however, significantly affect the binding of the different molecules to proteins and, consequently, NADH and NADPH perform separate functions within living cells. The redox potential for the NAD(P)þ/NAD(P)H couple is 320 mV vs SHE at pH 7. Since the reaction involves two electron and one proton this potential shifts by 29.5 mV for each unit change in pH at 298 K.

NH FAD N

O

R e + H+

H3C

H N

H3C

N

O NH Semiquinone N

O

R e + H+

H3C

H N

H3C

N

21

O NH FADH2

R

N H

O

1.5.4 Hemes

Figure 1.27 The redox reaction of flavin adenine dinucleotide.

The heme redox centre (also spelt haem) is a porphyrin ring, comprising four pyrrole rings linked by methylene bridges, with a single Fe ion coordinated in the centre. The different heme types, a, b and c differ in the substitution

O H3C

N

H3C

N

NH N

flavin

O

CH2

ribitol

HC OH HC OH

NH2

HC OH

N

O

N

O

O P O P O O

FAD,flavinadeninedinucleotide

N

CH2 N

adenosine

O

O H

H

H

OH

H

H

O H3C

N

H3C

N

NH N

O

CH2 HC OH HC OH

FMN,flavinmononucleotide

HC OH CH2 O O P O O

Figure 1.28 The structures of flavin adenine dinucleotide, FAD, and flavin adenine mononucleotide, FMN.

22

Bioelectrochemistry

O NH2 N

NAD+

H

O

O H

H

OH

H OH O

N

PO2 N

O P O

O H

O H

H H OH

OH

pattern found around the porphyrin ring (Figure 1.30). Because these substituents are directly attached to the ring, they directly affect the redox potential of the central Fe3þ/2þ NH 2 couple. Consequently the redox potentials of different hemes can vary quite widely. N The cytochrome f found in the photosynthetic cytochrome bf complex is a c-type cytochrome, but has a very N different protein structure; cytochrome f is an elongated protein with a largely b-sheet secondary structure in contrast to the usual a-helical structure found for most c-type cytochromes [6]. 1.5.5 Iron–Sulfur Clusters

H+, 2e Hα O



NH 2

NADH

N H

O

O H

H

OH

H OH O

N

PO2 N

O P O

O

O H

H

H

OH

H OH

Iron–sulfur clusters contain iron atoms bonded to both inorganic sulfur atoms and sulfur atoms on cysteine residues of the associated protein (Figure 1.31). In some 2Fe2S clusters, so-called Rieske clusters as found in the cytochrome bf and CoQH2–cyt c reductase complexes, two NH2 of the cysteine ligands are replaced by histidine. The iron atoms within the clusters have formal oxidation states of N either þ2 or þ3, but in actual fact the charge is delocalised between the iron atoms within the cluster [72]. The clusters N thus function as multielectron redox centres able to pick up or release electrons one at a time. 1.5.6 Copper Centres

Copper occurs as a one-electron centre going between the Cuþ and Cu2þ states. In plastocyanin, the copper is coordinated by N and S ligands from histidine, cysteine H H O and methionine amino acid residues in an environment NH2 which is distorted towards tetrahedral geometry. This NH 2 helps to stabilise the Cuþ state relative to Cu2þ [73], so N NADPH that the redox potential of Cu2þ/þ in plastocyanin in O N O N þ370 mV vs NHE, whereas the corresponding value for H H PO2 the aquo copper ion is þ170 mV. In cytochrome c oxidase H H O N N OH OH the three Cu atoms are in different environments. The single O P O Cu atom (Cub) is directly associated with the heme group of O H H O the cytochrome a3 at the site of oxygen reaction, whereas the H H other two of Cu atoms (the Cua site) are in distorted OH O tetrahedral coordination, bridged by two cysteine thiolates O P OH and coordinated by either histidine and methionine residues O or by histidine and glutamate residues [39]. As a result, the two Cua atoms are in different coordination environments Figure 1.29 The structures of the b-nicotinamide adenine and have been suggested to form a mixed valence complex.

dinucleotide redox couple, NADþ/NADH, and b-nicotinamide adenine dinucleotide phosphate, NADPH. The two faces of the nicotinamide ring, and thus the hydrogens at the C4 position in the reduced forms, NADH and NADPH, are not equivalent. In the figure the a-face faces out from the page and the b-face faces into the page.

1.6 GOVERNING PRINCIPLES Having described the electron transport pathways involved in energy transduction in some detail and looked

Bioenergetics and Biological Electron Transport

CH3

HC OH

CHCH 2

H3C N

N

heme a

Fe N

N CH3

OHC

O

O

O

O

CH3

CHCH 2

CHCH2

H 3C N

N Fe

N

heme b N CH3

H3C

O

O

O

(protein)RS

O

CH 3

H3C N

heme c N CH 3

H3C

O

O

respect to the lipid membrane is important, and that it is essential to ensure that the electron transfer reactions which occur within the electron transport chains occur between specific partners within the chain. If this breaks down, the organism will either not be able to capture energy efficiently from sunlight or will not be able to utilise the energy from food to make ATP. Thus, compounds which block the photosynthetic electron transport chain or intercept the mitochondrial electron transport chain are highly toxic to living systems. Defects in the efficiency of proton pumping across the mitochondrial membrane are associated with a wide range of human diseases, particularly those affecting the brain and muscle, where large amounts of ATP are used, although single organs or combinations of organs can be affected [18,19]. It is also clear that the properties of the phospholipid membrane, as a barrier to transport of protons between the inside and outside of the structure and as an environment in which to embed the large electron transfer proteins, is crucial in order to establish the proton motive force which ultimately drives synthesis of ATP in the living cell. We have also seen that these processes are achieved with a relatively restricted palette of redox centres and it is therefore of importance to consider the ways in which the properties of these redox centres, in particular their redox potentials, can be tuned within the system to optimise them for their place within the electron transport chain. We also need to consider the factors which control the rates of electron transfer between the different members of the electron transfer chain, both within multicentre redox proteins [74] and between different components. In this section we consider these different general points in turn starting with the membrane itself.

SR(protein)

N Fe

N

23

O

O

Figure 1.30 The structures of common hemes.

at the different redox centres involved, we are in a position to think about some of the common general guiding principles that determine the efficient operation of these systems. For example, it is clear from our discussion that organisation of the redox species with

1.6.1 Spatial Separation Phospholipids are amphiphilic molecules with a hydrophilic headgroup and a hydrophobic tail. To make the phospholipid membrane, the individual phospholipid molecules assemble with their headgroups on the outside and the hydrophobic tails on the inside in a bilayer membrane (Figure 1.32). This membrane is about 4 or 5 nm thick and is essentially impermeable to ions, including protons. The various membrane-bound proteins associated with the phospholipid bilayer make up a significant component of the overall membrane. For example, the inner mitochondrial membrane is typically 24 % lipid and 76 % protein, while the chloroplast membrane is only 25–30 % lipid and 70–75 % protein [75]. The lipid component is made up of a complex mixture of many different lipids that is different for

24

Bioelectrochemistry

H N

Cys-S

S-Cys S

Fe

N

His

Fe

S

S-Cys S Fe

Fe S

Cys-S

S-Cys

His

N

S-Cys

N H

Fe2S2

Rieske complex S-Cys

S Cys-S

Fe

Cys-S S

S Fe

S

Cys-S

Fe

Cys-S

Fe

S

S

Fe

S Fe

S

Fe

S-Cys

Fe3S4

S-Cys

Fe4S4

Figure 1.31 The structures of the common iron–sulfur clusters.

the mitochondrial membrane, the chloroplast and other biological membranes, and which is different for the inner and outer layers of the bilayer. This complexity, there are over 1000 different lipids in mammalian cells, indicates that there is some tuning of the properties of the lipid bilayer for different applications [76–78]. Figure 1.33 shows the structures of some common glycerophospholipids, that is phospholipids based on phosphatidic acid (3-sn-phosphatidic acid) esterified with different headgroups. Galactophospholipids make up as much as 70 % of the lipids in the chloroplast thylakoid membrane. Studies of the phospholipid composition of chloroplasts and mitochondria from avocado and cauliflower showed differences in the precise composition between the species and

major differences between the chloroplast and mitochondrion in both cases [76]. Diphosphatidylglycerol (cardiolipin) is highly enriched in the inner mitochondrial membrane and is not generally present in other cellular membranes; phosphatidylethanolamine is relatively more abundant and phosphatidylserine relatively less abundant in mitochondrial inner membranes [77]. The particular composition of the phospholipid bilayer determines its fluidity, which then regulates the properties of proteins embedded within it. It is also important to note that the composition of the lipid layer will not be homogeneous; there is good evidence for a non-uniform distribution of the different phospholipids within the membrane and association between specific proteins and specific phospholipids.

Figure 1.32 A schematic of the phospholipid bilayer membrane.

Bioenergetics and Biological Electron Transport

O O O

O P O O O

NH3 phosphatidylethanolamine

O P O O O

NH3

25

O O O O

O

phosphatidylserine

O

O O O O

O P O O O

NMe3

phosphatidylcholine

O O O O

O P O O O

OH OH

phosphatidylglycerol

O O O O

O

O O P

O O

OH O

O O

diphosphatidylglycerol (cardiolipin)

P O O

O O R1 R2

O O O

galactolipid O

O OH

CH2OH OH OH

Figure 1.33 The structures of some common phospholipids.

1.6.2 Energetics: Redox Potentials When we look at the energetics of electron transfer along the mitochondrial electron transport chain, from NADH at one terminus to oxygen at the other (Figure 1.34), we see

that the redox potentials of the different couples are organised in a steadily increasing sequence. At each step some part of the available free energy is used to drive the kinetics of electron transfer along the chain, while the remainder is used to pump protons across the mitochondrial membrane

26

Bioelectrochemistry

Figure 1.34 The sequence of redox potentials in the mitochondrial electron transport chain from NADþ to O2. The bars represent the range of potentials corresponding to the ratio of oxidised to reduced form from 1 : 10 to 10 : 1 (adapted from Smith and Wood [70]).

against the proton motive force (Figure 1.35). The sequence of redox potentials (Figure 1.34) also reveals the kinetic barrier for the reduction of oxygen to water, with by far the greatest drop in redox potential, 0.2 V, occurring in the final step between Cub/heme a3 of cytochrome c oxidase and the water/oxygen couple. This is a situation which is mirrored in current attempts to produce efficient fuel cells, where the slow electrode kinetics of the oxygen reduction reaction at neutral or acidic pH are a significant limitation on the efficiency of current fuel cells and biofuel cells. An examination of the sequence of redox potentials for the components of the photosynthetic electron transfer pathway reveals the same story, albeit in this case there are two large endergonic steps corresponding to the adsorption of two photons (Figure 1.36). This precise tuning of the redox potentials of the different constituents of the mitochondrial and photosynthetic electron transfer chains is achieved by control of the coordination sphere and environment of the different constituent redox centres. Thus, in the case of the heme proteins, the heme centre is bound to the polypeptide by two thioether bonds involving two cysteine residues. The redox potential of the FeIII/FeII centre in the heme is

altered by the coordination of two axial ligands, either a histidine and a methionine or two histidines above and below the heme plane, to the Fe centre and by the interaction of the heme with the surrounding polypeptide [79]. If we consider the full range of hemes in biological systems we find that the potentials for the Fe(III)/Fe(II) couple span 0.7 V from 0.3 V in histidine/histidine ligated heme c to þ0.4 V vs SHE in histidine/methionine ligated heme c [73]. This wide variation of redox potentials allows heme redox centres to fulfil roles at different stages all the way along the redox chain. In class I cytochromes c, where the two axial ligands are a histidine and a methionine, the redox potential of the Fe centre varies from þ0.2 to þ 0.38 V vs SHE [80] and this is attributed to the p electron withdrawing effect of the sulfur atoms of the thioether linkages and the axially bound methionine, all of which stabilise the Fe(II) state, and the poor solvent accessibility of the heme within the hydrophobic polypeptide pocket, again favouring the less charged Fe(II) state over the Fe(III) state. The final fine tuning of the redox potential is caused by changes in the electrostatic interactions between the charge on the Fe centre and the charges on polar amino acid residues within, and on the

Bioenergetics and Biological Electron Transport

Figure 1.35 The redox potentials for the three main complexes of the mitochondrial electron transport chain showing the number of protons pumped across the membrane at each stage for the transfer of two electrons along the chain. I is the NADH–CoQ reductase complex, II the CoQH2–Cyt c reductase complex and IV is the Cyt c oxidase complex; Q is ubiquinone and cyt c is cytochrome c.

surface of, the protein. For example, site directed mutagenesis studies of myoglobin have shown that replacing a valine (an uncharged amino acid residue) which is in van der Waals contact with the heme by glutamate or aspartate (both negatively charged amino acid residues) shifts the potential of the Fe(III)/Fe(II) couple by 0.2 V. Replacing the same valine by asparagine (which is uncharged) shifts the redox potential by 0.08 V [81]. In general, for type I cytochromes c, the modification of internal charges causes a 50 to 60 mV shift, whereas modification of surface charges has a smaller effect (10 to 30 mV) [79,80]. Similar effects have been demonstrated for cytochrome b562 variants [82]. A similar situation pertains for the flavin two electron, two proton redox couple. The redox potential of free flavin in solution at pH 7 is 0.21 V vs SHE [71]. However, the redox potentials of flavin in redox proteins and redox enzymes spans a wide range. In this case, the redox potential of the flavin is modulated by its immediate environment and by the significant difference between the typical dielectric constant for the interior of the protein, 5, and the much larger value for water, 78. Stacking interactions with aromatic amino acid residues above and below the flavin ring within the protein, also exert a

27

Figure 1.36 The sequence of redox potentials in the photosynthetic electron transport chain (adapted from Blankenship [6]).

significant effect on the redox potential [83–87]. In the case of flavodoxins, these effects lead to a large shift (around 0.45 V) of the flavin potential relative to that in water. The blue copper proteins are another example where the potentials of the redox centre span a range of values [88], in this case from 0.37 V vs NHE in P. nigra plastocyanin to 0.785 V vs. NHE in Polyporus versicolor laccase. Again, these differences can be explained by differences in the axial ligands around the Cu centre and by the degree of hydrophobicity of the polypeptide surrounding the redox centre. It is interesting to note that in the case of the blue copper proteins the coordination geometry about the Cu centre is virtually identical in the Cu(I) and Cu (II) redox states. Consequently there is very little reorganisation accompanying the electron transfer and the reorganisation energy (see below) is low, ER  0.6 to 0.8 eV, leading to fast electron transfer kinetics.

1.6.3 Kinetics: Electron Transfer Rate Constants Organising the redox potentials of the individual couples in the electron transport pathways in sequence is only one part of the story. The ordering of the redox potentials controls the thermodynamic driving force for the reaction between each sequential pair of components in the chain,

28

Bioelectrochemistry

but it cannot prevent non-specific electron transfer reactions between disparate components; indeed the thermodynamic driving force for these undesirable reactions may be significantly greater than for the specific, sequential electron transfer. Therefore, in addition to ordering of the redox potentials of the couples, there has to be significant selectivity in the kinetics of the reactions. This is achieved in several ways: spatial separation of components across the membrane, control of the distance for electron transfer, and the use of one- and two- electron couples. Electron transfer is, at its core, a quantum process (for an in depth discussion of the theory of electron transfer see [89]). The basic model for electron transfer reactions was developed by Marcus, Hush, Levich and Dogonadze in a series of contributions starting in the 1950s [90,91]. The basic versions of the theory are predicated on the separation of fast electronic motion and slow nuclear motion – the Franck–Condon principle. According to this model, electron transfer occurs by tunnelling between reactant and product nuclear vibrational surfaces. The activation energy for the process arises from the reorganisation of the nuclear and solvent coordinates required to bring the reactant system to a configuration in which the electron can transfer adiabatically to the product surface. In the basic treatment of electron transfer kinetics, the reactant and product energy surfaces are treated as parabolic and we have a model, as shown in Figure 1.37. Here

Figure 1.37 Gibbs free energy surface G(q) where q is the reaction coordinate, for an electron transfer reaction showing the reactant and product curves. DG0 is the thermodynamic driving force for the reaction, DGz is the activation free energy for the reaction, ER the reorganisation energy, qR, qz and qP are the configurations of the reactant, transition state and product, respectively.

q represents the reaction coordinate and includes both the inner and outer contributions, where the inner contributions refer to the nuclear rearrangement of the nuclei of the redox centres themselves and the outer contributions refer to the solvent environment around the reactant centres. DG0 is the thermodynamic driving force and DGz is the activation energy for the reaction. The rate constant for electron transfer between the donor and acceptor D þ A ! Dþ þ A

ð1:23Þ

is given by k DA ¼ kel

  weff  DGz exp RT 2p

ð1:24Þ

In this equation, oeff is the effective frequency of the nuclear motions which brings the system to the transition state configuration qz, and kel is the electron transfer probability at the nuclear configuration of the transition state. Assuming that the reactant and product energy surfaces are parabolic and of the same shape the free energy of activation can be written DGz ¼

ðE R  DG0 Þ2 4ER

ð1:25Þ

Equation (1.25) makes explicit the relationship between the activation energy and the thermodynamic driving force for the reaction, and predicts a quadratic dependence of ln(kDA) on DG0. This prediction has been verified for a number of systems including studies of electron transfer in ruthenium modified myoglobin [92], in which ruthenium redox centres were attached to a specific histidine residue on the surface of the protein and the rate of electron transfer between different ruthenium amine complexes and the active site were measured. In these experiments the thermodynamic driving force for the reaction was changed by changing the metal in the porphyrin at the active site (native Fe, Cd, Mg, Zn, Pd and proton) and by changing the ligands on the ruthenium attached to the peripheral histidine. In this way, the driving force for the electron transfer was varied from 0.39 to 1.17 eV giving around a 1000-fold change in the rate of intramolecular electron transfer. For the photosynthetic reaction centre, electron transfer has been shown to follow a quantum corrected Marcus expression [93]. The quantity ER in Equation (1.25) is the reorganisation energy. It corresponds to the free energy required to reor-

Bioenergetics and Biological Electron Transport

ganise the nuclear coordinates for the reactant from those corresponding to equilibrium for the reactants qR, to those corresponding to the equilibrium organisation for the products, qP, but without transfer of the electron (Figure 1.37). The reorganisation energy, ER, includes both the inner and outer sphere contributions to the reorganisation process. From Figure 1.37 it is clear that small values of ER, corresponding to similar structures for reactants and products, and/or broad parabolic free energy surfaces, corresponding to an easy reorganisation process, will give larger values of the electron transfer rate constant, kDA. In general, even for fast electron transfer reactions, ER is a large quantity many times bigger that kT. For biological systems, the protein surrounding the redox centre plays an important role in reducing the reorganisation energy and, as a consequence, speeding up electron transfer reactions [94]. A large part of this reduction in ER is caused by the exclusion of water. Bulk water has a high dielectric constant (e ¼ 78) and therefore interacts strongly with the charge, and the change in charge that accompanies electron transfer, at the redox centre. In contrast, the dielectric constant of the protein is lower (around five) and the interactions with the redox centre consequently smaller. Added to this, the constrained structure of the protein around the redox centre reduces the inner sphere contribution to ER. For example, the reorganisation energies for self-exchange in redox proteins are typically of the order of 0.6 to 0.8 eV [94], whereas the typical values for simple complex ions in aqueous solution are significantly larger (around 2 eV). Although the reorganisation process is important in determining the kinetics of electron transfer between donor and acceptor, in order to determine the origin of the selectivity in the kinetics of electron transfer between specific partners in the biological electron transfer pathways, we must turn our attention to the pre-exponential term kel in Equation (1.24). According to the semi-classical model kel is given by rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi p kel ¼ ð1:26Þ H2 2 h ER RT DA where H2DA is the electronic coupling matrix element and describes the strength of coupling between the reactant and product states at the nuclear configuration of the transition state. If we assume that the electron transfer occurs by tunnelling through a square, uniform barrier using the Hopfield model, then    bðr  r0 Þ H DA ðrÞ ¼ H DA ðr 0 Þexp 2

ð1:27Þ

29

where r0 is the van der Waals contact distance for the donor and acceptor and b describes the exponential attenuation of the overlap with distance between the donor and acceptor. The simple exponential dependence of the rate of electron transfer with distance predicted by Equations (1.24) to (1.27) can be understood in terms of the drop off of the electronic wavefunctions of the donor HOMO and acceptor LUMO orbitals with distance. In general, these will decrease exponentially at larger distances from the redox centre; the overlap of these two exponentially decaying wavefunctions, then, itself yields an exponential dependence on the separation of donor and acceptor. There is some debate as to whether the distance between donor and acceptor should be measured between the metal centres in the redox groups in species such as hemes, or between the edges of the ligands [94]. The parameter b in Equation (1.27) depends on the nature of the intervening medium through which the electron tunnels. For an electron tunnelling through a vacuum, b is generally taken to be between 30 and 50 nm1 [94]. However, when the electron tunnels through an intervening medium between the donor and acceptor, the value of b is reduced, because the height of the tunnel barrier is reduced; the electron is able to travel further. For protein a value of b of 14 nm1 has been suggested based on a range of experimental measurement [93]. This means that for every 0.17 nm increase in the distance between the donor and acceptor, the rate of electron transfer will decrease by a factor of 10. The topic of long range electron transfer remains an area of significant interest, not only within bioelectrochemistry, but also in the field of molecular electronics and nanotechnology [95,96] and there have been many elegant studies on model systems. One of these that is particularly relevant to bioelectrochemistry is the work of Isied et al. [97] who studied the distance dependence of electron transfer between two metal complexes separated by an oligoproline bridge (Figure 1.38). Oligoproline bridges were selected because their rigidity allows the spacing between donor and acceptor to be well defined; the molecule cannot fold up to bring the donor and acceptor ends close to each other. They found an exponential decrease in the rate of electron transfer with increasing distance up to three or four proline molecules in the bridge (corresponding to a donor/acceptor spacing of 2 nm), and beyond that a levelling off of the electron transfer rate. They suggest several possible explanations for this change in terms of different pathways for electron transfer, including the possibility of a change from through-space to through-bond electron transfer.

30

Bioelectrochemistry

O 0

(NH3)5Os N O Ru(NH3)5 O O

1

(NH3)5Os N

O Ru(NH3)5 N

O O 2

(NH3)5Os N

N N O

O Ru(NH3)5

O O 3

(NH3)5Os N

N

O

N

O Ru(NH3)5 O

N

O

O O 4

N

(NH3)5Os N

O Ru(NH3)5

O

N

N O

N

Figure 1.38 The structures of a set of oligoproline bridged ruthenium–osmium donor–acceptor complexes (based on Isied et al. [97]).

The study of electron transfer in proteins has been significantly advanced over the last 15 years by our increasing knowledge of the precise three-dimensional structures of various redox proteins and by the application of protein engineering to allow the systematic investigation of the effects of the mutation of individual amino acids on the rates of electron transfer. The simple Hopfield model described above treats the protein between the donor and acceptor sites as a homogeneous, featureless medium. In practice, the donor and acceptor are separated by the peptide, which comprises various amino acid residues in a particular sequence and orientation. This has led to a vigorous debate as to whether there exist specific electron transfer pathways within proteins [94,98– 102], or whether the process can be described purely in terms of a distance dependence [103,104]. In the tunnelling pathway model proposed by Beratan et al. [98,99,105], the pathway between the donor and acceptor is divided into a number of segments corresponding to covalently bonded parts, hydrogen bonded parts, and through-space parts

where there is van der Waals contact. The overall decay of the tunnelling interaction is then treated as the product of the decay across each block. There is then an approximately exponential decay in coupling with the number of blocks in the tunnelling pathway, so that relatively few pathways make an important contribution to the coupling between donor and acceptor within the model. These contributing pathways are identified using a structure searching algorithm. Support for this model comes from studies of electron transfer in ruthenium modified cytochromes c [100], where better agreement is found between the logarithm of the electron transfer rate and the calculated path length rather than the physical distance. According to the pathway model, the coupling through a b peptide strand should be greater than that through an a helix, because in the a helix the distance between the two ends increases much more slowly, and non-linearly, with the helix length, as compared to the case for the more linear b strand [102,106] (Figure 1.39). Using site directed mutagenesis, Langen et al. attached ruthenium

Bioenergetics and Biological Electron Transport

H N O OH O

O NH

O NH O

NH O H N

HN

O

HO

O

H N O

alpha helix NH2

HN O

O

O

NH NH O

N H

31

O

H N O

N H

O

H N O

N H

O

H N O

N H

H N

NH2 O

beta strand Figure 1.39 The structures for the a helix and b strand showing the relative lengths of the two for a decapeptide.

complexes to histidine residues introduced at different places along the b barrel of the azurin from Pseudomonas aeruginosa, and measured the rate of electron transfer between the ruthenium and the Cu(I) [102]. They found an exponential dependence of the rate of electron transfer with distance, with a decay constant of 11.0 nm1, close to that predicted for coupling along a b strand by the pathway model. The experimental studies described above refer to intramolecular electron transfer in proteins. In the biological system, intermolecular electron transfer also plays a significant role. Here again we can expect the distance dependence to play an important role. In the case of intermolecular electron transfer, particularly when one or both of the reactants is large, the orientation of the two molecules will have a significant effect and allows selectivity to be introduced into the process – electron transfer will be fast between those components which bind together in the correct orientation to bring the redox centres close together. However, from the point of view of experimental studies, the uncertainty in the precise geometry of the complex formed between the reactants makes it much harder to study the fine details of electron transfer in the reaction complex formed by two large proteins. Recent interest has focused on the role of water molecules between the two reactants and the existence of electron tunnelling pathways through the structured water trapped between the proteins [107]. 1.6.4 Size of Proteins A consequence of the need for selectivity in electron transfer reactions in electron transport chains, and therefore the concomitant need to control the environment

around the redox centre to tune its potential and the distance dependence of the rate of electron transfer, is that the redox proteins involved have to be reasonably large molecules. For example, if we consider the multicentre transmembrane proteins, such as the NADH–CoQ reductase complex or photosystem II, it is clear that they must be large enough to span the phospholipid membrane and that, in addition, they need to be larger enough to separate out the different redox centres within the protein in order to control the rates of electron transfer between the different centres. In essence, a separation of more that about 1.4 nm is required in order to kinetically limit electron transfer between centres [104]. The possible factors which determine the size of soluble proteins have been described by Goodsell and Olson [108]. They highlight three effects. First, the requirement that the length of the poly(peptide) chain is long enough to enforce the overall shape of the protein [109]. This is supported by the observation that many smaller proteins contain disulfide linkages or rigid metal clusters which enhance the stability of the folded protein. Second, the suggestion, originally attributed to Pauling [108], that in order to control association between proteins it is necessary to decrease the ratio of surface to volume and to ensure that the surface area is large enough to allow patterning of the surface, so that there can be specific interactions between reaction partners [109]. This effect is seen, for example, in the interaction of plastocyanin with cytochrome f in the chloroplast or the interaction between cytochrome c and the cytochrome bc1 complex in the mitochondrion. Finally, Goodsell and Olson suggest that, for some reactions, surface diffusion, where the reactant diffuses across the protein surface to the active site, is an important factor [110].

32

Bioelectrochemistry

1.6.5 One-Electron and Two-Electron Couples Looking back at the mitochondrial or photosynthetic electron transport chains (Figures 1.8 and 1.18), it is clear that a final factor which provides control and selectivity over electron transfer is the division between one-electron redox couples, such as hemes or Cu centres, and two-electron redox couples such as NADH. For the NAD(P)/NAD(P)H couple, the one-electron oxidation or reduction intermediate is not readily accessible and redox reactions in biological systems proceed by a formal hydride transfer mechanism. As a consequence, the rates of reaction between NAD(P)H and one-electron couples in the electron transfer chain, although thermodynamically favourable, are slow. This, together with the fact that oxygen has a triplet ground state, also accounts for the slow reaction of NAD(P)H with oxygen, another factor which is vital for the operation of both mitochondrial and photosynthetic electron transport chains. Thus, the NAD (P)H couple is able to exchange electron pairs with redox couples such as succinate/fumarate or lactate/pyruvate, but not to short circuit electron transfer directly to oxygen or other components of the electron transport chain. Clearly the two-electron redox world of NAD(P)H and the one-electron redox world of the hemes, iron–sulfur clusters and Cu proteins need to be linked together and this is achieved by the flavin redox couple which is equally at home as a two-electron redox couple reacting with NAD(P)H or as a redox couple reacting in two oneelectron steps via a thermodynamically and kinetically viable semiquinone radical intermediate [71]. Examples of this linkage can be found in the crystallographic studies of ferredoxin-NADPþ reductase [66,111,112] and glutathione reductase [113,114]. These show that the hydride transfer occurs through a stacked flavin–NADPþ complex that confers the correct geometry for the hydride transfer (Figure 1.40).

Figure 1.40 The stacked flavin NADPþ structure required for hydride transfer based on the structure in the binding site of ferredoxin reductase (based on Karplus and Bruns [111]). Ser 92 and Cys 272 are residues of the ferredoxin involved in binding the flavin and NADPþ. The aromatic rings of the flavin and NADPþ are held face to face 0.33 nm apart with the nicotinamide C4 opposite the flavin N5 atom.

1.7 ATP SYNTHASE It is appropriate that we conclude this chapter with a discussion of ATP synthase. The ATP synthase enzymes from different organisms, including those from mitochondria, chloroplasts, fungi and bacteria, show a very high degree of conservation in structure and function. The enzyme sits in the membrane and utilises the proton motive force generated by the different electron transport chains described above to carry out the synthesis of ATP from ADP and inorganic phosphate. ATP synthase [5,115–118] (Figure 1.41), is a multisubunit enzyme made up of an F0 membrane bound portion

Figure 1.41 The general structure of F0F1 ATP synthase from Escherichia coli (adapted from Capaldi and Aggeler [5]). The a, b, three a, three b, and d sub-units make up the stator and the nine to twelve c, g and e sub-units make up the rotator.

Bioenergetics and Biological Electron Transport

and a soluble F1 portion which can be dissociated by treatment in low ionic strength buffer. This soluble F1 portion remains catalytically active after dissociation. The F0, membrane bound, portion of ATP synthase from Escherichia coli is made up of three sub-units, a, b and c, in the ratio 1 : 2 : 10–14. The F0 portion provides a specific proton conduction channel between the c sub-unit ring and the a unit. The catalytic sites for ATP synthesis are located in the F1 portion of the enzyme, which comprises three subunits, a, b and g, in the ratio 3 : 3 : 1. The a and b sub-units form a hexagon with the three catalytic sites located at the interface between the units. The helical region of the g subunit passes through the core of the a3b3 hexagon. The other end of the g sub-unit and the e sub-unit are securely attached to the c ring. During the catalytic cycle, the passage of protons through the F0 potion of the enzyme drives the c ring around and thus causes the helical portion of the g sub-unit to rotate within the a3b3 hexagon, which is itself held stationary by viscous drag from the two b sub-units and the d sub-unit. In the catalytic cycle, the three catalytic sites pass sequentially through three different conformations – substrate binding, formation of tightly bound ATP, release of ATP – brought about by the rotation of the g sub-unit (Figure 1.42). Each 120 rotation leads to the synthesis, or

Figure 1.42 The proposed mechanism for ATP synthesis (based on Capaldi and Aggeler [5] and Yashuda et al. [126]). The three catalytic sites (each ab sub-unit pair) are in different conformations; one is open (b) and ready to bind ADP and inorganic phosphate or ATP, one is partly open (bADP) and contains ADP and inorganic phosphate, the third is closed (bATP) and contains ATP. Rotation of the central g sub-unit drives the catalytic sites sequentially through the sequence leading to the endergonic conversion of ADP and inorganic phosphate to ATP.

33

hydrolysis, of one molecule of ATP. Thus in ATP synthase, mechanical rotation is the mechanism by which the free energy of the proton motive fore is converted into the chemical potential of ATP. Evidence for this mechanism comes from a variety of sources, including radio labelling and specific cross-linking studies [119,120], fluorescent labelling studies [121], by nmr [122], by cross-linking [123] and through direct observation of single molecules using epi-fluorescent microscopy [124–126]. In these experiments, Yasuda et al. immobilised F1-ATPase from E. coli onto Ni–NTA modified surfaces by histidine tagging of the a and b subunits. The rotation of the g sub-unit was then observed by attaching a biotinylated, fluorescently labelled actin filament, around 5 nm in length, through strepavidin to the g sub-unit. When 2 mM ATP was added to the solution, they observed the continuous rotation of the actin filament in an anti-clockwise direction when viewed from above (corresponding to the membrane side) with a rotational torque of 40 pN nm1. In subsequent experiments [125], they showed that at low ATP concentration, 12 h at 80  C). This wired H2O2 sensor showed good thermostability, and the current decreased at a rate of myoglobin (Mb) > KatG > catalase. KatG catalyzed the electrochemical reduction of oxygen more efficiently than catalase and CcP, but less efficiently than the other peroxidases. DMPC films incorporating glucose oxidase and peroxidases gave good analytical responses to glucose, demonstrating the feasibility of functional dual enzyme–lipid films [256]. The molybdo-enzyme dimethylsulfoxide (DMSO) reductase from Rhodobacter capsulatus gave MoVI/MoV and MoV/MoIV peaks in DDAB films on edge plane PG electrodes [257]. Catalytic activity was observed by voltammetry in the presence of DMSO. DDAB films on edge plane PG were also used to obtain voltammetry of porcine purple acid phosphatase (Uteroferin). The one-electron reversible CV peaks in this non-heme di-iron enzyme were attributed to FeIII–FeIII/FeIII–FeII, and binding of phosphate and arsenate inhibitors was investigated [258]. Reaction center (RC) enzymes from photosynthetic purple bacterium Rhodobacter sphaeroides and the photosystem I (PSI) RC from spinach, contain multiple bound redox cofactors. Thin films of DMPC and the bacterial RC at 4  C revealed a reproducible, chemically irreversible oxidation peak at 0.98 V and a reduction peak at 0.17 V vs NHE [259]. Disappearance of this reduction peak when the quinones were removed from these bacterial RCs, suggesting that the peak at 0.17 corresponded to reduction of quinone cofactors. The peak at 0.98 V decreased by 85%

when illuminated by visible light. Furthermore, this peak gave a catalytic response in solutions of ferrous cytochrome c, the natural redox partner of the bacterial RC. These results suggested assignment of the 0.98 V peak to oxidation of the primary electron donor of the bacterial RC enzyme. On gold electrodes, bacterial RC–lipid film voltammetry was complicated by halide ion adsorption, which gave interfering peaks that complicated interpretation [260]. Electron transfer between PG electrodes and native spinach PS I RC gave two well-defined chemically reversible reduction–oxidation peaks in DMPC films. These peaks were assigned to phylloquinone, A1 (Em ¼ 0.54 V) and iron–sulfur clusters FA/FB (Em ¼ 0.19 V) by comparisons with PS I samples selectively depleted of these cofactors [166]. As mentioned earlier, analysis of CV data by the Butler–Volmer thin-film model gave ks values 7.2 s1 for A1 and 65 s1 for FA/FB. A catalytic process was observed in which electrons were injected from terminal FA/FB cofactors of PS I in the films, to iron–sulfur protein ferredoxin in solution, mimicking in vivo electron shuttle during photosynthesis. 2.3.7 Polyion Films and Layer-by-Layer Methods Films of enzymes and polymers have been utilized often for mediated enzyme electrochemistry in sensors. However, cast or spin coated polymer films have seen only moderate use for direct electron transfer studies of enzymes. Films of polyelectrolytes support high water content when in contact with electrolyte solutions and can serve as suitable hosts for proteins. Poly(lysine) has protonated amine groups at neutral pH, and can facilitate coadsorption of negatively charged proteins on electrodes. For example, proton reduction catalyzed by a negative hydrogenase was studied by using poly(lysine) on a PG electrode [261]. Poly(lysine) was added to solutions to achieve direct electron transfer from indium tin oxide (ITO) electrodes to spinach ferredoxin, which then transferred electrons to

72

Bioelectrochemistry

cyt P450cam in solution. As mentioned earlier, poly(lysine), spinach ferredoxin and cyt P450cam were used in this fashion to dehalogenate toxic organohalides [151]. Although cast polymer–protein films are easy to make, the protein may leak from the films and CV peaks can decay with time [10]. In any case, this approach has not been used extensively for enzyme electrochemistry. Layer-by-layer construction of enzyme–polyion films provides excellent control over thickness, and allows film architecture to be designed according to the plans of the builder. Film thickness and enzyme loading can be increased by adding more layers, and several different enzymes can be incorporated into a film. Alternate adsorption of monolayers of biomolecules and polyions has been developed for various applications by Lvov, Decher and others [47–52,262].

Layer-by-layer electrostatic film assembly on a negatively charged electrode is illustrated in Figure 2.17. In this example, we start with a negatively charged electrode that can be obtained by oxidation of carbon, by treating a metal oxide with base, or by using an organothiol SAM terminating in sulfonate or carboxylate on gold or silver. This negatively charged electrode is then immersed into a 1–3 mg mL1 solution of positively charged polyions. At this concentration, the polycations adsorb at steady state in about 15–20 min [51,52], effectively reversing the charge on the outer surface. In some cases, better enzyme loading is obtained by using several initial layers of oppositely charged polyions. In the example in Figure 2.17, after rinsing with water, the electrode is immersed in a 1–3 mg mL1 solution of negatively charged enzyme. A negative surface charge on the

Figure 2.17 Construction of an enzyme film by using layer-by-layer alternate electrostatic assembly of the enzyme with polyions.

Electrochemistry of Redox Enzymes

73

Some ionic polymers used for film formation Polycations

H n

H

n

+ N

N+ H3 C

poly(ethylene imine) (PEI)

CH3

poly(diallydimethylamine) (PDDA)

Polyanions

n

DNA

SO3 – poly(styrenesulfonate) (PSS)

enzyme is selected by using a buffer of pH larger than the isoelectric point. After the enzyme is adsorbed, the outer surface develops a negative charge. The adsorption steps are repeated as many times as desired to obtain a multilayer assembly. For an initial electrode charge that is positive, the procedure would be revised so that the charges of the polyions and enzyme in Figure 2.17 are reversed. Film growth can be monitored during or after each adsorption step with quartz crystal microbalance (QCM) weighing, surface plasmon resonance, spectroscopy, voltammetry or other suitable methods. The bottom of Figure 2.17 depicts the final assembled film. Nearest neighbor protein and polyion layers are intimately mixed. Neutron reflectivity studies of films of polycations and polyanions, and PSS and the protein myoglobin, confirmed extensive mixing of neighboring layers. However, on the smooth silicon surfaces used in those studies, appearance of neutron Bragg peaks for well-separated deuterated PSS layers suggested the first and third (or fourth) layers were spatially distinct in the films [50–52]. Most enzymes retain excellent activity and near native conformations in these films. Some polyions that can be used to make layer-by-layer enzyme films are shown above, and include metal oxide nanoparticles and DNA. The layer-by-layer method was first used for enzyme electrochemistry with cyt P450cam to obtain direct electron exchange with gold electrodes [263]. Films



nanoparticles: TiO2

MnO2

SiO2

Clay

required an undercoating SAM of mercaptopropanesulfonate (MPS) on gold to facilitate direct voltammetry of the enzymes. This SAM also placed negative sulfonate groups at the electrode-solution interface to adsorb the first layer of polycation or cationic protein. Chemically reversible cyclic voltammograms were found for the FeIII/FeII redox couple of cyt P450cam on smooth gold. Films containing cyt P450cam or myoglobin on smooth vapor-deposited gold electrodes had only about 1.3 to 1.5 electroactive layers, as shown by comparing total enzyme measured by QCM with electroactive enzyme from voltammetry [263]. The number of electroactive layers can be greatly increased by using mechanically roughened PG electrodes, which may provide a disorder-inducing template that enhances electron transport by ‘electron hopping’ between enzyme redox sites [264]. Also, when polyions were deposited from solutions with relatively high salt concentrations, where they are coiled rather than linear, much more protein was adsorbed in the subsequent deposition step. Films constructed on rough PG electrodes gave up to seven electroactive protein layers. Films were stable for several months upon storage dry or in buffer solutions at 4  C. Reversible voltammetry was also found for films of polycations and putidaredoxin, the natural ferredoxin redox partner of bacterial cyt P450cam [265]. The Eo dispersion model gave a good fit to SWV data giving

74

Bioelectrochemistry

average ks of 4.5 s1. Unfortunately, reduced putidaredoxin in the polyion films did not feed electrons to its natural redox partner cyt P450cam in the films, because of unfavorable Eo shifts of the enzymes in the film. QCM and atomic force microscopy (AFM) were used to characterize binding of cyt P450s to layers of polyions as well as binding to protein redox partners [54]. AFM showed that cationic PEI and anionic PSS as initial layers on smooth metal surfaces formed polymer islands featuring globular structures of 10 nm in diameter. By the time a fourth alternate layer of PEI or PSS had been adsorbed, the polymer islands merged and enzyme adsorption as the fifth layer took place on a nearly continuous, relatively smooth surface. Cyt P450cam and cyt P450 2B4 appeared as 10 nm globules on the polymer underlayers by tapping mode AFM. This study showed that the cyt P450s, which have non-uniform charge distributions with negative and positive ‘patches’ on the ends of the macromolecule, could be oriented with either the positively or negatively charged end up, by controlling the charge sign of the underlying polyion layer. Consequently, cyt P450s can be adsorbed to both negative and positive polyion layers for voltammetry. Evidence from electrochemical studies with other proteins suggests that such ‘bivalent’ binding of proteins may be a relatively general phenomenon [266]. Enzyme-catalyzed epoxidation of styrene and its derivatives catalyzed by cyt P450cam was compared in surfactant

and layer-by-layer polyion films [250,263]. In this reaction, a catalytic electrochemical reduction drives an enzymecatalyzed oxidation in a doubly catalytic process. The overall pathway (Scheme 2.14; P ¼ protein) is similar to that first suggested for myoglobin (Mb) in aqueous solutions and microemulsions [267]. The electrode converts PFeIII into PFeII, which reacts rapidly with dioxygen to give PFeII–O2. PFeII–O2 is reduced to give H2O2, which converts MbFeIII to active oxidant radical PFeIV¼O. This ferryloxy radical epoxidizes styrene by oxygen transfer to the double bond. Cyclic voltammetry in the presence of oxygen gave a large reduction peak at the PFeIII potential of 0.35 vs SCE at pH 7.4, but addition of styrene caused only a slight increase in the catalytic wave, and no kinetic data could be obtained. Electrolyses at 0.6 V vs SCE at 4  C in oxygenated pH 7.4 buffer saturated with styrene or cismethylstyrene, showed that enzyme-polyion films with two cyt P450cam layers on Au–MPS electrodes gave the largest turnover rates. Improved catalysis of styrene epoxidation by enzyme–polyion films was related to better mechanical stability of these films compared to enzyme-surfactant films. Stereochemistry for the epoxidation of cis-b-methylstyrene by cyt P450cam–polyion films depended on oxygen availability, suggesting two competing pathways [250]. A stereoselective pathway utilizes the ferryloxy radical (Scheme 2.15), as suggested for the natural enzyme system. A non-stereoselective pathway may involve a peroxy radical

+e – +

H + PF e

III

-e



PFe

II

O2

II

PFe –O 2 –

2e , 2H +

H2 O2 •PFe

IV

H2O2 + PFe

=O + H 2 O

II

active oxidant

o III

PFe + Scheme 2.14 Electrochemical peroxide dependent epoxidation of styrene catalyzed by cyt.

Electrochemistry of Redox Enzymes

75

o CH3

·PFe CH3

IV

cis =O

O2

·O-O-PFe IV = O o

o

CH3

CH3 +

cis

trans

Scheme 2.15 Suggested oxygen-dependent stereoselecivity in epoxidation pathways catalyzed by cyt P450 and Mb.

on the enzyme surface that forms by reaction of the ferryloxy radical with dioxygen. A later study addressed optimization of catalytic and electrochemical properties of cyt P450cam films constructed with alternate layers on rough PG electrodes with respect to film thickness, polyion type and pH [268]. Alternate layers of polyions such as poly(styrene sulfonate) (PSS), as opposed to SiO2 nanoparticles or DNA, supported the best catalytic and electrochemical performance. Charge transport involving the iron heme proteins was achieved through 40–320 nm in these films, and probably involves electron hopping facilitated by interlayer mixing. However, very thin films (c. 12–25 nm) gave the largest turnover rates for the catalytic epoxidation of styrene, while thicker films showed reactant transport limitations. Classical bell-shaped activity–pH profiles and turnover rates similar to those in solution suggested that films grown layer-by-layer are suitable for turnover rate studies of redox enzymes. Recombinant human cyt P450 3A4 films were also assembled on gold electrodes by alternate adsorption with layers of polycations. Direct, reversible electron transfer between the gold electrode and cyt P450 3A4 was observed with cyclic and square-wave voltammetry under anaerobic conditions [269]. In the presence of oxygen, catalytic reduction peaks were observed. Addition of the cyt P450 3A4 substrates verapamil, midazolam, quinidine and progesterone to the oxygenated buffer caused a concentration-dependent increase in current for the catalytic reduction of oxygen. Product analyses after electrolysis with the enzyme film confirmed the formation of the expected metabolites. Furthermore, metabolite formation was inhibited by ketoconazole, a known inhibitor of cyt P450 3A4. Results suggested that thin films of cyt P450s on electrodes are applicable to drug screening applications.

Films of recombinant human cyt P450 1A2 and PSS were constructed on carbon cloth electrodes using the layer-by-layer alternate absorption method and evaluated for electrochemical- and H2O2-driven enzyme-catalyzed oxidation of styrene to styrene oxide. With electrochemical initiation, epoxidation of styrene was mediated by initial catalytic reduction of dioxygen to H2O2 which activates the enzyme for the catalytic oxidation (see Scheme 2.14) [270]. Slightly larger turnover rates for cyt P450 1A2 were found for electrolytic-and H2O2driven reactions compared to conventional enzymatic reactions using cyt P450s, their reductases, and electron donors for cytochrome P450 1A2. Cyt P450cam gave comparable turnover rates in film electrolysis and solution reactions. Results showed that cyt P450 1A2 catalyzes styrene epoxidation faster than cyt P450cam.

2.4 OUTLOOK FOR THE FUTURE In this chapter, we have traced the growth of enzyme electrochemistry from early stages in the 1970s and early 1980s, when fundamental electrochemical studies of enzymes were difficult at best and often impossible. Research progress into the 21st century has provided a wealth of excellent tools for electrochemical studies of enzymes at high levels of mechanistic detail. Important advances include theory for mediated and direct electron transfer as well as enzyme catalysis in films, and the development of an array of film construction methods providing the possibility to study direct electron transfer and catalysis with electrochemical methods for virtually any redox enzyme. It appears to us that the statement ‘electron transfer (involving a given enzyme) is slow because the redox site is buried within the enzyme

76

Bioelectrochemistry

structure’ has been overused in the past. Even glucose oxidase, which has been mediated for applications to commercial glucose sensors, has fallen prey to direct electron transfer using special wiring techniques or carbon nanotube electrodes. Advances in enzyme electrochemistry have resulted in an increased use of thin-film voltammetry for studies of enzymes, and had led to the elucidation of some rather complex and unusual gated catalytic mechanisms that would be difficult to study with other methods. However, at the time of this writing, general simulation and modeling software for thin-film voltammetry do not seem to be readily available to general users. Commercialization of such packages similar to those on offer for analysis of CV and SWV solution electrochemistry [271] would be of great value. Nevertheless, the future appears bright for enzyme electrochemistry, and we look forward to the eventual appearance of voltammetry in the everyday toolboxes of most enzyme biologists.

ACKNOWLEDGEMENTS Financial support of JFR’s research program from grant no. ES03154 from the National Institute of Environmental Health Sciences, NIH, grant no. CTS-0335345 from NSF, grant no. DAAD-02-1-0381 from the US Army Research Office and grant no. 2002-35318-12484 US Department of Agriculture (USDA) is gratefully acknowledged. JR is also indebted to students and colleagues named in joint publications without which none of the work from his laboratory would have been possible.

REFERENCES 1. J. D. Czaban, Electrochemical sensors in clinical chemistry: yesterday, today, and tomorrow, Anal. Chem., 57, 345A–356A (1985). 2. G. Ramsay (ed.), Commerical Biosensors. Wiley, New York, 1998. 3. E. F. Bowden, F. M. Hawkridge and H. N. Blount, Electrochemical aspects of bioenergetics, in S. Srinivasan, Y. A. Chizmadzhev, J. O’ M. Bockris, B. E. Conway and E. Yeager (eds.), Comprehensive Treatise of Electrochemistry, Vol. 10, Plenum, New York, 1985, pp. 297–346. 4. F. A. Armstrong, Probing metalloproteins by voltammetry, in Bioinorganic Chemistry, Structure and Bonding 72, Springer-Verlag, Berlin, 1990, pp. 137–221. 5. G. G. Guilbault and M. Mascini, Analytical Uses of Immobilized Biological Compounds for Detection, Medical, and Industrial Uses, Reidel Pub. Co., Holland, 1988.

6. J. M. Kauffmann and G. G. Guilbault, Enyzme electrode biosensors: theory and applications, in Bioanalytical Applications of Enzymes, Methods of Biochemical Analysis Vol. 36, Wiley, NY, 1992, pp. 63–113. 7. G. G. Guilbault, Analytical Uses of Immobilized Enzymes. Marcel Dekker, New York, 1984. 8. P. R. Ortiz de Montellano and C. E. Catalano, Epoxidation of styrene by hemoglobin and myoglobin, J. Biol. Chem., 260, 9265–9271 (1985). 9. C. E. Castro and E. W. Bartnicki, Conformational isomerism and effective redox geometry in the oxidation of heme proteins by alkyl halides, cyt c and cyt oxidase, Biochemistry, 14, 498–502 (1975). 10. J. F. Rusling and Z. Zhang, Thin films on electrodes for direct protein electron transfer, in R. W. Nalwa (ed.), Handbook of Surfaces and Interfaces of Materials, Vol. 5. Biomolecules, Biointerfaces, and Applications. Academic Press, San Diego, 2001, pp. 33–71. 11. J. F. Rusling and Z. Zhang, Polyion and surfactant films on electrodes for protein electrochemistry, in J. Q. Chambers and A. Brajter-Toth (eds.), Electroanalytical Methods for Biological Materials, Marcel Dekker, New York, 2002, pp. 195–231. 12. F. A. Armstrong, H. A. Heering and J. Hirst, Reactions of complex metalloproteins studies by protein film voltammetry, J. Chem. Soc. Rev., 26, 169–179 (1997). 13. P. N. Bartlett, P. Tebbutt and R. G. Whitaker, Kinetic aspects of modified electrodes and mediators in bioelectrochemistry, Prog. Reaction Kinetics, 16, 55–155 (1991). 14. A. Cass, G. Davis, G. D. Francis et al. Ferrocene-mediated enzyme electrode for amperometric determination of glucose, Anal. Chem., 56, 667–671 (1984). 15. F. A. Armstrong, H. A. O. Hill and N. J. Walton, Direct electrochemistry of redox proteins, Acc. Chem. Res., 21, 407–413 (1988). 16. E. F. Bowden, F. M. Hawkridge, J. C. Chlebowski, E. E. Bancroft, C. Thorpe and H. N. Blount, Cyclic voltammetry and derivative cyclic voltabsorptivity of purified horse heart cytochrome c at tin-doped indium oxide optically trasnparent electrodes, J. Am. Chem. Soc., 104, 7641–7644 (1982). 17. D. E. Reed and F. M. Hawkridge, Direct electron transfer reactions of cytochrome c at silver electrodes, Anal. Chem., 59, 2334–2339 (1987). 18. T. Yokota, K. Itoh and A. Fujishima, Redox behavior of cytochrome c immobilized into a lecithin monolayer and deposited onto SnO2 by the Langmuir–Blodgett technique, J. Electroanal. Chem., 216, 289–292 (1987). 19. Z. Salamon and G. Tollin, Reduction of cytochrome c at a lipid bilayer electrode, Bioelectrochem. Bioenerg., 25, 447–454 (1991). 20. Z. Salamon and G. Tollin, Interfacial electrochemistry of cytochrome c at a lipid bilayer modified electrode: effect of incorporation of negative charges into the bilayer on cyclic voltammetric parameters, Bioelectrochem. Bioenerg., 26, 321–334 (1991).

Electrochemistry of Redox Enzymes

21. M. J. Tarlov and E. F. Bowden, Electron transfer reaction of cytochrome c adsorbed on carboxylic acid terminated alkanethiol monolayer electrodes, J. Am. Chem. Soc., 113, 1847–1849 (1991). 22. R. A. Marcus and N. Sutin, Electron transfers in chemistry and biology, Biochim. Biophys. Acta, 811, 265–322 (1985). 23. G. McLendon, Long-distance electron transfer in proteins and model systems, Acc. Chem. Res., 21, 160–167 (1988). 24. M. Dequaire, B. Limoges, J. Moiroux and J. M. Saveant, Mediated electrochemistry of horseradish peroxidase. Catalysis and inhibition, J. Am. Chem. Soc., 124, 240–253 (2002). 25. B. Limoges, J. M. Saveant and D. Yazidi, Quantitative analysis of catalysis and inhibition at horseradish peroxidase monolayers immobilized on an electrode surface, J. Am. Chem. Soc., 125, 9192–9203 (2003). 26. V. Niviere, E. C. Hatchikian, P. Bianco and J. Haladjian, Kinetic studies of electron transfer between hydrogenase and cytochrome c3 from Desulfovibrio gigas, Biochim. Biophys. Acta, 935, 34–40 (1988). 27. C. Moreno, R. Franco, I. Moura, J.Le Gall and M. J. J. G. Moura Voltammetric studies of the catalytic electrontransfer process between the Desulfovibrio gigas hydrogenase and small proteins isolated from the same genus, Eur. J. Biochem., 217, 981–989 (1993). 28. A. Heller, Electrical wiring of redox enzymes, Acc. Chem. Res., 23, 128–134 (1990). 29. A. Heller, Electrical connection of enzyme redox centers to electrodes, J. Phys. Chem., 96, 3579–3587 (1992). 30. J. F. Rusling, R. J. Forster, Electrochemical catalysis with redox polymer and polyion-protein films, J. Colloid Inter. Sci., 262, 1–15 (2003). 31. B. A. Greg and A. Heller, Redox polymer films containing enzymes. 1. A redox-conducting epoxy cement: synthesis, characterization, and electrocatalytic oxidation of hydroquinone, J. Phys. Chem., 95, 5970–5975 (1991). 32. B. A. Greg and A. Heller, Redox polymer films containing enzymes. 2. Glucose oxidase containing enzyme electrodes, J. Phys. Chem., 95, 5976–5980 (1991). 33. I. Katakis and A. Heller, L-b-glycerophosphate and L-lactate electrodes based on the electrochemical ‘wiring’ of oxidases, Anal. Chem., 64, 1008–1013 (1992). 34. T. J. Ohara, R. Rajagopalan and A. Heller, Glucose electrodes based on cross-linked bis-(2,20 -bipyridine) chloroosmiumþ /2 þ complexed poly(1-vinylimidazole) films, Anal. Chem., 65, 3512–3517 (1993). 35. T. J. Ohara, R. Rajagopalan and A. Heller, ‘Wired’ enzyme electrodes for amperometric determination of glucose or lactate in the presence of interfering substances, Anal. Chem., 66, 2451–2457 (1994). 36. C. Taylor, G. Kenausis, I. Katakis and A. Heller, ‘Wiring’ of glucose oxidase within a hydrogel made with polyvinyl imidazole complexed with [(Os-4,40 -dimethoxy-2,20 -bipyridine)Cl]2 þ /1 þ J. Electroanal. Chem, 396, 511–515 (1995).

77

37. F. Mao, N. Mano and A. Heller, Long tethers binding redox centers to polymer backbones enhance electron transport in enzyme ‘wiring’ hydrogels, J. Am. Chem. Soc., 125, 4951–4957 (2003). 38. N. Mano, F. Mao and A. Heller, A miniature biofuel cell operating in a physiological buffer, J. Am. Chem. Soc., 124, 12962–12963 (2002). 39. T. Tatsuma, K. Saito and N. Oyama, Enzyme electrodes mediated by a thermoshrinking redox polymer, Anal. Chem., 66, 1002–1006 (1994). 40. M. Vreeke, R. Maidan and A. Heller, Hydrogen peroxide and ß-nicotinamide adenine dinucleotide sensing amperometric electrodes based on electrical connection of horseradish peroxidase redox centers to electrodes through a three-dimensional electron relaying polymer network, Anal. Chem., 64, 3084–3090 (1992). 41. M. S. Vreeke, K. T. Yong and A. Heller, A thermostable hydrogen peroxide sensor based on ‘Wiring’ of soybean peroxidase, Anal. Chem., 67, 4247–4249 (1995). 42. A. Badia, R. Carlini, A. Fernandez, F. Battaglini, S. R. Mikkelsen and A. M. English, Intramolecular electrontransfer rates in ferrocene-derivatized glucose oxidase, J. Am. Chem. Soc., 115, 7053–7060 (1993). 43. B. A. Gregg and A. Heller, Cross-linked redox gels containing glucose oxidase for amperometric biosensor applications, Anal. Chem., 62, 258–263 (1990). 44. J. J. Calvente, A. Narvaez, E. Domiguez and R. Andreu, Kinetic analysis of wired enzyme electrodes: application to horseradish peroxidase entrapped in a redox polymer matrix, J. Phys. Chem. B., 107, 6629–6643 (2003). 45. N. Anicet, A. Anne, J. Moiroux and J. M. Saveant, Electron transfer in organized assemblies of biomolecules. Construction and dynamics of avidin/biotin co-immobilized glucose oxidase/ferrocene monolayer carbon electrodes, J. Am. Chem. Soc., 120, 7115–7116 (1998). 46. J. Hodak, R. Etchenique, E. J. Calvo, K. Singhal and P. N. Bartlett, Layer-by-layer self-assembly of glucose oxidase with a poly(allylamine)ferrocene redox mediator, Langmuir, 13, 2708–2716 (1997). 47. Y. Lvov, G. Decher and G. Sukhorukov, Assembly of thin films by means of successive deposition of alternate layers of DNA and poly(allylamine), Macromolecules, 26, 5396–5399 (1993). 48. C. Bourdillon, C. Demaille, J. Moiroux and J. M. Saveant, From homogeneous electroenzymatic kinetics to antigenantibody construction and characterization of spatially ordered catalytic enzyme assemblies on electrodes, Acc. Chem. Res., 29, 529–535 (1996). 49. C. Demaille, J. Moiroux, J. M. Saveant and C. Bourdillon, Anitgen–antibody assembling of enzyme monomolecular layers and multimonolayers on electrodes, in Y. Lvov and H. M€ohwald (eds.), Protein Architecture: Interfacing Molecular Assemblies and Immobilization Biotechnology. Marcel Dekker, New York, 2000, pp. 311–335.

78

Bioelectrochemistry

50. G. Decher, Fuzzy nanoassemblies, Science, 277, 1231–1237 (1997). 51. Y. Lvov, Electrostatic layer-by-layer assembly of proteins and polyions, in Y. Lvov and H. M€ohwald (eds.), Protein Architecture: Interfacing Molecular Assemblies and Immobilization Biotechnology. Marcel Dekker, New York, 2000, pp. 125–167. 52. Y. Lvov, Thin film nanofabrication by alternate adsorption of polyions, nanoparticles and proteins, in R. W. Nalwa (ed.), Handbook of Surfaces and Interfaces of Materials, Vol. 3. Nanostructured Materials, Micelles and Colloids, Academic Press, San Diego, 2001, pp. 170–189. 53. J. F. Rusling and R. J. Forster, Electrochemical catalysis with redox polymer and polyion-protein films, J. Colloid Inter. Sci., 262, 1–15 (2003). 54. J. B. Schenkman, I. Jansson, Y. Lvov, J. F. Rusling, S. Boussaad and N. J. Tao, Charge-dependent sidedness of cytochrome P450 forms studied by QCM and AFM, Archives Biochem. Biophys., 385, 78–87 (2001). 55. E. J. Calvo, R. Etchenique, L. Pietrasanta, A. Wolosiuk and C. Danilowicz, Layer-by-layer self-assembly of glucose oxidase and Os(bpy)2ClPyCH2NH-poly(allylamine) bioelectrode, Anal. Chem., 73, 1161–1168 (2001). 56. V. Rosca and I. C. Popescu, Kinetic analysis of horseradish peroxidase ‘wiring’ in redox polyelectrolyte-peroxidase multilayer assemblies, Electrochem. Commun., 4, 904–911 (2002). 57. E. J. Calvo and A. Wolosiuk, Supramolecular architectures of electrostatic self-assembledglucose oxidase enzyme electrodes, Chem. Phys. Chem., 5, 235–239 (2004). 58. S. B. Adeloju and G. G. Wallace, Conducting polymers and the bioanalytical sciences: new tools for biomolecular communications: a review, Analyst, 121, 699–703 (1996). 59. P. N. Bartlett and J. M. Cooper, A review of the immobilization of enzymes in electropolymerized films, J. Electroanal. Chem., 362, 1–12 (1993). 60. M. V. Deshpande and D. P. Amalnerkar, Biosensors prepared from electrochemically-synthesized conducting polymers, Prog. Polym. Sci., 18, 623–649 (1993). 61. P. N. Bartlett, Z. Ali and V. Eastwickfield, Electrochemical immobilisation of enzymes. Part 4. Co-immobilisation of glucose oxidase and ferro/ferricyanide in poly(N-methyl pyrrole) films, J. Chem. Soc. Faraday Trans., 88, 2677–2683 (1992). 62. C. Iwakura, Y. Kajiya and H. Yoneyama, Simultaneous immobilization of glucose oxidase and a mediator in conducting polymer films, J. Chem. Soc. Chem. Commun., 1019–1020 (1988). 63. Y. Kajiya, H. Sugai, C. Iwakura and H. Yoneyama, Glucose sensitivity of polypyrrole films containing immobilized glucose oxidase and hydroquinonesulfonate ions, Anal. Chem., 63, 49–54 (1991). 64. W. Schuhmann, Electron-transfer pathways in amperometric biosensors. Ferrocene-modified enzymes entrapped in

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

conducting-polymer layers, Biosens. Bioelectron., 10, 181–193 (1995). S. Cosnier, A. Deronzier and J. C. Moutet, Oxidative electropolymerization of polypyridinyl complexes of ruthenium(II)-containing pyrrole groups, J. Electroanal. Chem., 193, 193–204 (1985). W. Schuhmann, C. Kranz, J. Huber and H. Wohlschlager, Conducting polymer-based amperometric enzyme electrodes. Towards the development of miniaturized reagentless biosensors, Synth. Met., 61, 31–35 (1993). N. C. Foulds and C. R. Lowe, Immobilization of glucose oxidase in ferrocene-modified pyrrole polymers, Anal. Chem., 60, 2473–2478 (1998). K. Habermuller, A. Ramanavicius, V. Laurinavicius and W. Schuhmann, An oxygen-insensitive reagentless glucose biosensor based on osmium-complex modified polypyrrole, Electroanalysis, 12, 1383–1389 (2000). S. Reiter, K. Eckhard, A. Blochl and W. Schuhmann, Redox modification of proteins using sequential-parallel electrochemistry in microtiter plates, Analyst, 126, 1912–1918 (2001). W. Schuhmann, Conducting polymer based amperometric enzyme electrodes, Mickrochim. Acta, 121, 1–29 (1995). M. Aizawa, S. Yabuki, H. Shinohara and Y. Ikariyama, Electrically regulated biocatalytic processes of redox enzymes embedded in conducting polymer membrane, Ann. N. Y. Acad. Sci., 613, 827–831 (1990). S. Yabuki, H. Shinohara and M. Aizawa, Electro-conductive enzyme membrane, J. Chem. Soc., Chem. Commun., 945–946 (1989). P. D. Poet, S. Miyanoto, T. Murakami, J. Kimura and I. Karube, Direct electron transfer with glucose oxidase immobilized in an electropolymerized poly(N-methylpyrrole) film on a gold microelectrode, Anal. Chim. Acta, 235, 255–263 (1990). C. J. Koopal, B. D. Ruite and R. J. M. Nolte, Amperometric biosensor based on direct communication between glucose oxidase and a conducting polymer inside the pores of a filtration membrane, J. Chem. Soc., Chem. Commun., 1691–1692 (1991). C. J. Koopal, M. C. Fciters, R. J. M. Nolte, B. D. Ruiter and R. B. M. Schasfoort, Glucose sensor utilizing polypyrrole incorporated in tract-etch membranes as the mediator, Biosen. Bioelectron., 7, 461–471 (1992). S. Kuwabata and C. R. Martin, Mechanism of the amperometric response of a proposed glucose sensor based on a polypyrrole-tubule-impregnated membrane, Anal. Chem., 66, 2757–2762 (1994). T. Ruzgas, A. Lindgren, L. Gorton et al. Electrochemistry of peroxidases, in J. Q. Chambers and A. Brajter-Toth (eds.), Electroanalytical Methods for Biological Materials, Marcel Dekker, New York, 2002, pp. 233–275. U. Wollenberger, V. Bogdanovskaya, S. Bobrin, F. Scheller and M. Tarasevich, Enzyme electrodes using bioelectro-

Electrochemistry of Redox Enzymes

79.

80.

81.

82.

83.

84.

85.

86.

87.

88. 89.

90.

91.

catalytic reduction of hydrogen peroxide, Anal. Lett., 23, 1795–1808 (1990). T. Tatsuma, M. Gondaira and T. Watanabe, Peroxidaseincorporated polypyrrole membrane electrodes, Anal. Chem., 64, 1183–1187 (1992). T. Tatsuma and T. Watanabe, Electrochemical characterization of polypyrrole bienzyme electrodes with glucose oxidase and peroxidase, J. Electroanal. Chem., 356, 245–253 (1993). Y. Yang and S. Mu, Bioelectrochemical responses of the polyaniline horseradish peroxidase electrodes, J. Electroanal. Chem., 432, 71–78 (1997). (a) B. Lu, M. R. Smyth and R. O’Kennedy, Immunological activities of IgG antibody on precoated Fc receptor surfaces, Anal. Chim. Acta, 331, 97–102 (1996). (b) A. J. Killard, S. Zhang, H. Zhao, R. John, E. I. Iwuoha and M. R. Smyth, Development of an electrochemical flow injection immunoassay (FIIA) for the real time monitoring of biospecific interactions, Anal. Chim. Acta, 400, 109–119 (1999). (c) A. J. Killard, L. Micheli, K. Grennan et al. Anal. Chim. Acta, 427, 173–180 (1999). T. Tatsuma, T. Ogawa, R. Sato and N. Oyama, Peroxidaseincorporated sulfonated polyaniline–polycation complexes for electrochemical sensing of H2O2, J. Electroanal. Chem., 501, 180–185 (2001). X. Yu, G. A. Sotzing, F. Papadimitrakepoulos and J. F. Rusling, Wiring of enzymes to electrodes by ultrathin conductive polyion underlayers: enhanced catalytic response to hydrogen peroxide, Anal. Chem., 75, 4565–4571 (2003). A. Ramanavicius, K. Habermuller, E. Csoregi, V. Laurinavicius and W. Schuhmann, Polypyrrole-entrapped quinohemoprotein alcohol dehydrogenase. evidence for direct electron transfer via conducting-polymer chains, Anal. Chem., 71, 3581–3586 (1999). G. F. Khan, H. Shnohara, Y. Ikariyama and M. Aizawa, Electrochemical behaviour of monolayer quinoprotein adsorbed on the electrode surface, J. Electroanal. Chem., 315, 263–273 (1991). T. Ikeda, S. Miyaoka, F. Matsushi, D. Kobayashi and M. Senda, Direct bioelectrocatalysis at metal and carbon electrodes modified with adsorbed D-gluconate dehydrogenase or adsorbed alcohol dehydrogenase from bacterial membranes, Chem. Lett., 847–850 (1992). P. N. Bartlett and Y. Astier, Microelectrochemical enzyme transistors, Chem. Comm., 105–112. (2000). H. S. White, G. P. Kittlesen and M. S. Wrighton, Chemical derivatization of an array of three gold microelectrodes with polypyrrole: fabrication of a molecule-based transistor, J. Am. Chem. Soc., 106, 5375–5377 (1984). T. Mastue, M. Nishizawa, T. Sawaguchi and I. Uchida, An enzyme switch sensitive to NADH, J. Chem. Soc., Chem. Commun., 1029–1030 (1991). D. T. Hoa, T. N. S. Kumar, N. D. Punekar, et al. A biosensor based on conducting polymers, Anal. Chem., 64, 2645–2646 (1992).

79

92. P. N. Bartlett and P. R. Birkin, Enzyme switch responsive to glucose, Anal. Chem., 65, 1118–1119 (1993). 93. P. N. Bartlett and P. R. Birkin, A microelectrochemical enzyme transistor responsive to glucose, Anal. Chem., 66, 1552–1559 (1994). 94. K. Seelbach and U. Krag, Nanofiltration membranes for cofactor retention in continuous enzymatic synthesis, Enzyme Microb. Technol., 20, 389–392 (1997). 95. V. Kitpreechavanich, N. Nishio, M. Hayashi and S. Nagai, Regeneration and retention of NADP(H) for xylitol production in an ionized membrane reactor, Biotechnol. Lett., 7, 657–662 (1985). 96. P. Gacesa and J. Hubble, Enzyme Technology. Taylor & Francis, New York, 1987. 97. P. Gacesa and R. F. Vennio, The preparation of stable enzyme–coenzyme complexes with endogenous catalytic activity, Biochem. J., 177, 369–372 (1979). 98. H. K. Chenault and G. M. Whitesides, Regeneration of nicotinamide cofactors for use in organic synthesis, Appl. Biochem. Biotechnol., 14, 147–197 (1987). 99. W. Hummel, Large-scale application of NAD(P)-dependent oxidoreductases: recent developments, Trends in Biotechnol., 17, 487–492 (1999). 100. J.-P. Vandecasteele, Enzymatic synthesis of L-carnitine by reduction of an achiral precursor: the problems of reduced nicotinamide adenine dinucleotide recycling, Appl. Environ. Microbiol., 39, 327–334 (1980). 101. D. Mandler and I. Willner, Photosensitized NAD(P)H regeneration systems, J. Chem. Soc. Perkin Trans., 2, 805–811 (1986). 102. Z. Goren, N. Lapidot and I. Willner, Photocatalyzed regeneration of NAD(P)H by CdS and TiO2 semiconductors: Applications in enzymatic synthesis, J. Mol. Catal., 47, 21–23 (1988). 103. I. Willner and D. Mandler, Enzyme-catalysed biotransformations through photochemical regeneration of nicotinamide cofactors, Enzyme Microb. Technol., 11, 467–483 (1989). 104. H. Jaegfeldt, A study of the products formed in the electrochemical reduction of nicotinamide-adenine-dinucleotide, Bioelectrochem. Bioenerg., 8, 355–370 (1981). 105. M. Aizawa, R. W. Coughlin and M. Charles, Electrolytic regeneration of the reduced from the oxidized form of immobilized NAD, Biotechnol. Bioeng., 28, 209–215 (1976). 106. M. Aizawa, S. Suzuki and M. Kubo, Electrolytic regeneration of NADH from NAD þ with a liquid crystal membrane electrode, Biochim. Biophys. Acta, 444, 886–892 (1976). 107. Y.-T. Long and H.-Y. Chen, Electrochemical regeneration of coenzyme NADH on a histidine modified silver electrode, J. Electroanal. Chem., 440, 239–242 (1997). 108. R. DiCosimo, C.-H. Wong, L. Daniels and G. M. Whitesides, Enzyme-catalyzed organic synthesis: electrochemical regeneration of NAD(P)H from NAD(P) using methyl viologen and flavoenzymes, J. Org. Chem., 46, 4622–4623 (1981).

80

Bioelectrochemistry

109. J. C. Hoogvliet, L. C. Lievense, C. V. Dijk and C. Veeger, Electron transfer between the hydrogenase from Desulfovibrio vulgaris (Hildenborough) and viologens, Eur. J. Biochem., 174, 273–280 (1988). 110. R. Ruppert, S. Herrmann and E. Steckhan, Efficient indirect electrochemical insitu regeneration of NADH: electrochemically driven enzymatic reduction of pyruvate catalyzed by DLDH, Tetrahedron Lett., 28, 6583–6586 (1987). 111. G. Hilt and E. Steckhan, Transition metal complexes of 1,10-phenanthroline-5,6-dione as efficient mediators for the regeneration of NAD þ in enzymatic synthesis, J. Chem. Soc. Chem. Commun., 1706–1707 (1993). 112. M. D. Leonida, S. B. Sovolov and A. J. Fry, FAD-mediated enzymatic conversion of NAD þ to NADH: application to chiral synthesis of L-lactate, Bioorg. Med. Chem. Lett., 8, 2819–2824 (1998). 113. M. Kim and S. Yun, Construction of electro-enzymatic bioreactor for the production of (R)-mandelate from benzoylformate, Biotechnol. Lett., 26, 21–26 (2004). 114. M. D. Leonida, A. J. Fry, S. B. Sobolov and K. I. Voivodov, Co-electropolymerization of a viologen oligomer and lipoamide dehydrogenase on an electrode surface: application to cofactor regeneration, Bioorg. Med. Chem. Lett., 6, 1663–1666 (1996). 115. R. J. Fisher, J. M. Fenton and J. Iranmahboob, Electroenzymatic synthesis of lactate using electron transfer chain biomimetic membranes, J. Membr. Sci., 177, 17–24 (2000). 116. M. T. Grimes and D. G. Drueckhammer, Membraneenclosed electroenzymatic catalysis with a low molecular weight electron-transfer mediator, J. Org. Chem., 58, 6148–6150 (1993). 117. Z. Shaked and G. M. Whitesides, Enzyme-catalyzed organic synthesis: NADH regeneration by using formate dehydrogenase, J. Am. Chem. Soc., 102, 7104–7105 (1980). 118. C.-H. Wong, D. G. Drueckhammer and H. M. Sweers, Enzymatic vs fermentative synthesis: thermostable glucose dehydrogenase catalyzed regeneration of NAD(P)H for use in enzymatic synthesis, J. Am. Chem. Soc., 107, 4028–4031 (1985). 119. C.-H. Wong and G. M. Whitesides, Enzyme-catalyzed organic synthesis: NAD(P)H cofactor regeneration by using glucose 6-phosphate and the glucose-6-phosphate dehydrogenase from Leuconostoc mesenteroides, J. Am. Chem. Soc., 103, 4890–4899 (1981). 120. S. S. Godbole, S. F. D’Souza and G. B. Nadkami, Regeneration of NAD(H) by alcohol dehydrogenase in gel-trapped yeast cells, Enzyme Microb. Technol., 5, 125–128 (1983). 121. C.-H. Wong, L. Daniels, W. H. Orme-Johnson and G. M. Whitesides, Enzyme-catalyzed organic synthesis: NAD(P) H regeneration using dihydrogen and the hydrogenase of Methanobacterium thermoautotrophicum, J. Org. Chem., 103, 6627–6628 (1981). 122. B. Bossow and C. Wandrey, Continuous enzymically catalyzed production of L-leucine from the corresponding

123.

124.

125.

126.

127.

128.

129.

130.

131.

132.

133.

134.

135.

136.

racemic hydroxy acid, Ann. N.Y. Acad. Sci., 506, 631–636 (1987). C.-H. Wong and J. R. Matos, Enantioselective oxidation of 1,2-diols to L-a-hydroxy acids using coimmobilized alcohol and aldehyde dehydrogenase as catalyst, J. Org. Chem., 50, 1992–1994 (1985). A. Liese, M. Karutz, J. Kamphuis, C. Wandrey and U. Kragl, Enzymatic resolution of 1-phenyl-1,2-ethanediol by enantioselective oxidation: overcoming product inhibition by continuous extraction, Biotechnol. Bioeng., 51, 544–550 (1996). Y. Tsuji, T. Fukui, T. Kawamoto and A. Tanaka, Enantioselective dehydrogenation of b-hydroxysilanes by horse liver alcohol dehydrogenase with a novel in situ NADþ regeneration system, Appl. Microbiol. Biotechnol., 41, 219–224 (1994). D. G. Drueckhammer and C.-H. Wong, FMN reductase catalyzed regeneration of NAD(P) for use in enzymic synthesis, J. Org. Chem., 50, 5387–5389 (1985). S. Itoh, T. Terasaka, M. Matsumiya, M. Komatsu and Y. Ohshiro, Efficient NAD þ -recycling system for ADH-catalysed oxidation in organic media, J. Chem. Soc. Perkin Trans., 1, 3253–3254 (1992). H. R. Zare and S. M. Golabi, Electrocatalytic oxidation of reduced nicotinamide adenine dinucleotide (NADH) modified glassy carbon electrode, J. Electroanal. Chem., 464, 14–23 (1999). B. Grundig, G. Wittstock, U. Rupert and B. Strehlitz, Mediator-modified electrodes for electrocatalytic oxidation of NADH, J. Electroanal. Chem., 395, 143–157 (1995). H. Ju and D. Leech, [Os(bpy)2(PVI)10Cl]Cl polymer-modified carbon fiber electrodes for the electrocatalytic oxidation of NADH, Anal. Chim. Acta, 345, 51–58 (1997). A. Malinauskas, T. Ruzgas and L. Gorton, Electropolymerization of preadsorbed layers of some azine redox dyes on graphite, J. Coll. Interf. Sci., 224, 325–332 (2000). T. Osa, Y. Kashiwagi and Y. Yanagisawa, Electroenzymatic oxidation of alcohols on a poly(acrylic acid)-coated graphite felt electrode terimmobilizing ferrocene, diaphorase and alcohol dehydrogenase, Chem. Lett., 23, 367–368 (1994). M. Sadakane and E. Steckhan, Electrochemical properties of polyoxometalates as electrocatalysts, Chem. Rev., 98, 219–237 (1998). N. F. Atta, I. Marawi, K. L. Petticrew, H. Zimmer, H. B. Mark and A. Galal, Electrochemistry and detection of some organic and biological molecules at conducting polymer electrodes, J. Electroanal. Chem., 408, 47–52 (1996). R. Antiochia, I. Lavagnini and F. Magno, Electrocatalytic oxidation of dihydronicotinamide adenine dinucleotide with ferrocene carboxylic acid by diaphorase from Clostridium kluyveri. Remarks on the kinetic approaches usually adopted, Electroanal., 11, 129–133 (1999). D. Schwartz, M. Stein, K.-H. Schneider and F. Giffhorn, Synthesis of D-xylulose from D-arabitol by enzymatic conversion with immobilized mannitol dehydrogenase

Electrochemistry of Redox Enzymes

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.

149.

150.

151.

from Rhodobacter spaeroides, J. Biotechnol., 33, 95–101 (1994). Y. Ogino, K. Takagi, K. Kano and T. Ikeda, Reaction between diaphorase and quinone compounds in bioelectrocatalytic redox reactions of NADH and NAD þ , J. Elecroanal. Chem., 396, 517–524 (1995). K. Takagi, K. Kano and T. Ikeda, Mediated bioelectrocatalysis based on NAD-related enzymes with reversible characteristics, J. Electroanal., 445, 211–219 (1998). E. Simon and P. N. Bartlett, Modified electrodes for the oxidation of NADH, in Biomolecular Films, J. F. Rusling (ed.), Surfactant Science Series 111, Marcel Dekker, New York, 2003. A. Anne, C. Bourdillon, S. Daninos and J. Moiroux, Can the combination of electrochemical regeneration of NAD þ , selectivity of L-a -amino-acid dehydrogenase, and reductive amination of a-keto-acid be applied to the inversion of configuration of a L-a -amino-acid, Biotechnol. Bioeng., 64, 101–107 (1999). J. M. Obon, P. Casanova, A. Manjon, V. M. Femandez and J. L. Iborra, Stabilization of glucose dehydrogenase with polyethyleneimine in an electrochemical reactor with NAD(P) þ regeneration, Biotechnol. Prog., 13, 557–561 (1997). O. Miyawaki and T. Yano, Electrochemical bioreactor with regeneration of NAD þ by rotating graphite disk electrode with PMS adsorbed, Enzyme Microb. Technol., 14, 474–478 (1992). J. Handman, A. Harriman and G. Porter, Photochemical dehydrogenation of ethanol in dilute aqueous solution, Nature, 307, 534–535 (1984). M. Julliard and J. Le Petit, Regeneration of NAD þ and NADP þ cofactors by photosensitized electron transfer, Photochem. Photobiol., 36, 283–290 (1982). I. Willner, R. Maidan and B. Willner, Photochemically induced oxidative and reductive regeneration of NAD(P) þ / NAD(P)H cofactors: applications in biotransformations, Isr. J. Chem., 29, 289–301 (1989). A. Sharma and M. S. N. Quantrill, Measurement of ethanol using fluorescence quenching, Spectrochim. Acta (Part A), 50, 1161–1177 (1994). A. Sharma and M. S. N. Quantrill, a-Ketoglutarate assay on fluorescence quenching by NADH, Biotechnol. Prog., 12, 413–416 (1996). V. Repia, M. P. Mayhew and V. L. Vilker, A direct electrode-driven Cyt P450 cycle for biocatalysis, Proc. Natl. Acad. Sci. USA, 94, 13554–13558 (1997). M. P. Mayhew, V. Repia, M. J. Holden and V. L. Vilker, Improving the Cyt P450 enzyme system for electrodedriven biocatalysis of styrene epoxidation, Biotechnol. Prog., 16, 610–616 (2000). J. Kazlauskaite, A. C. G. Westlake L.-L. Wong and H. A. O. Hill, Direct electrochemistry of cytochrome P450cam, Chem. Commun., 2189–2190 (1996). M. Wirtz, J. Klucik and M. Rivera, Ferredoxin-mediated electrocatalytic dehalogenation of haloalkanes by Cyto-

152.

153.

154.

155. 156.

157.

158.

159.

160.

161.

162.

163.

164.

165.

81

chrome P450cam, J. Am. Chem. Soc., 122, 1047–1056 (2000). K. Niki and B. W. Gregory, Electrochemistry of redoxactive protein films immobilized on self-assembled monolayers of organothiols, in J. F. Rusling (ed.), Biomolecular Films, Marcel Dekker, New York, 2003, pp. 65–98. R. W. Murray, Chemically modified electrodes, in Electroanalytical Chemistry, A. J. Bard (ed.), Marcel Dekker, New York, 1984, Vol. 13, pp. 191–368. R. W. Murray, Introduction to the chemistry of molecularly designed electrodes, in R. W. Murray (ed.), Molecular Design of Electrode Surfaces, Techniques of Chemistry Series. Wiley-Interscience, New York, 1992, Vol. 22, pp. 1–48. A. J. Bard and L. R. Faulkner, Electrochemical Methods, 2nd edn., Wiley, New York, 2001. A. El Kasmi M. C. Leopold, R. Galligan et al. Adsorptive immobilization of cytochrome c on indium/tin oxide (ITO): electrochemical evidence for electron transfer-induced conformational changes, Electrochem. Commun., 4, 177–181 (2002). M. Madja, Dynamics of electron transport in polymeric assemblies of redox centers, in R. W. Murray (ed.), Molecular Design of Electrode Surfaces, Techniques of Chemistry Series. Wiley-Interscience, New York, 1992, Vol. 22, pp. 159–206. J. F. Rusling and Z. Zhang, Designing functional biomolecular films on electrodes, in J. F. Rusling (ed.), Biomolecular Films. Marcel Dekker, New York, 2003, pp. 1–64. Z. Zhang, S. Chouchane, R. S. Magliozzo and J. F. Rusling, Reversible electrochemistry and catalysis with mycobacterium tuberculosis catalase-peroxidase in lipid films, Chem. Commun., 177–178 (2001). T. M. Nahir, R. A. Clark and E. F. Bowden, Linear sweep voltammetry of irreversible electron transfer in surface confined species using the Marcus theory, Anal. Chem., 66, 2595–2598 (1994). Z. Zhang and J. F. Rusling, Electron transfer between myoglobin and electrodes in phosphatidyl choline and dihexadecyl phosphate films, Biophys. Chem., 63, 133–146 (1997). A.-E. F. Nassar Z. Zhang, N. Hu, J. F. Rusling and T. F. Kumosinski, Proton-coupled electron transfer from electrodes to myoglobin in ordered biomembrane-like films, J. Phys. Chem. B, 101, 2224–2231 (1997). F. A. Armstrong, Applications of voltammetric methods for probing the chemistry of redox proteins, in G. Lenz and G. Milazzo (eds.), Bioelectrochemistry of Biomacromolecules, Birkhauser Verlag, Basel, Switzerland, 1997, pp. 205–255. J. Hirst and F. A. Armstrong, Fast scan cyclic voltammetry of protein films of pyrolytic graphite edge electrodes: characteristics of electron exchange, Anal. Chem., 70, 5062–5071 (1998). E. Laviron, General expression of the linear potential sweep voltammogram in the case of diffusionless electrochemical systems, J. Electroanal. Chem., 101, 19–28 (1968).

82

Bioelectrochemistry

166. B. Munge, S. K. Das, R. Ilagan et al. Electron transfer reactions of redox cofactors in spinach photosystem I reaction center protein in lipid films on electrodes, J. Am. Chem. Soc., 125, 12457–12463 (2003). 167. N. Oyama and T. Ohsaka, Voltammetric diagnosis of charge transport on polymer coated electrodes, in R. W. Murray (ed.), Molecular Design of Electrode Surfaces, Techniques of Chemistry Series. Wiley-Interscience, New York, 1992, Vol. 22, pp. 333–402. 168. T. M. Nahir and E. F. Bowden, The distribution of standard rate constants for electron transfer between thiol-modified gold electrodes and adsprbed cytochrome c, J. Electroanal. Chem., 410, 9–13 (1996). 169. C. L. Leger, S. J. Elliot, K. R. Hoke et al. Enzyme electrokinetics: using protein film voltammetry to investigate redox enzymes and their mechanisms, Biochemistry, 42, 8653–8662 (2003). 170. A. K. Jones, R. Cambra, G. A. Reid, S. K. Chapman and F. A. Armstrong, Interruption and time-resolution of catalysis by a flavoenzyme using fast scan protein film voltammetry, J. Am. Chem. Soc., 122, 6494–6495 (2000). 171. A. Sucheta, R. Cammack, J. Weiner and F. A. Armstrong, Reversible voltammetry of fumarate reductase immobilized on an electrode surface, Biochemistry, 32, 5455–5465 (1993). 172. C. P. Andrieux and J. M. Saveant, Catalysis at redox polymer coated electrodes, in R. W. Murray (ed.), Molecular Design of Electrode Surfaces, Techniques of Chemistry Series. WileyInterscience, New York, 1992, Vol. 22, pp. 207–270. 173. K. Alcantera and J. F. Rusling, Voltammetric measurement of Michaelis-Menten kinetics for a protein in a lipid film reacting with a protein in solution, Electrochem. Comm., 7, 223–226 (2005). 174. M. J. Honeychurch and P. V. Bernhardt, A numerical approach to modeling the catalytic voltammetry of surface-confined redox enzymes, J. Phys.Chem. B, 108, 15900–15909 (2004). 175. J. Osteryoung and J. J. O’Dea, Square wave voltammetry, in A. J. Bard (ed.), Electroanalytical Chemistry, Vol. 14, Marcel Dekker, New York, 1986, pp. 209–308. 176. J. J. O’Dea and J. Osteryoung, Characterization of quasireversible surface processes by square wave voltammetry, Anal. Chem., 65, 3090–3097 (1993). 177. Z. Zhang, A-E. F. Nassar, Z. Lu, J. B. Schenkman and J. F. Rusling, Direct electron injection from electrodes to cytochrome P450cam in biomembrane-like films, J. Chem. Soc. Faraday Trans., 93, 1769–1774 (1997). 178. T. M. Saccucci and J. F. Rusling, Modeling square wave voltammetry of thin protein films using Marcus theory, J. Phys Chem. B, 105, 6142–6147 (2001). 179. L. J. C. Jeuken, A. K. Jones, S. K. Chapman, G. Cecchini and F. A. Armstrong, Electron-transfer mechanisms through biological redox chains in multicenter enzymes, J. Am. Chem. Soc., 124, 5702–5713 (2002). 180. L. J. C. Jeuken, J. P. McEvoy and F. A. Armstrong, Insights into gated electron-transfer kinetics at the electrode protein interface: a square wave voltammetry study of the blue

181.

182.

183.

184.

185.

186.

187.

188.

189.

190.

191.

192.

193.

194.

copper protein azurin, J. Phys Chem. B, 106, 2304–2313 (2002). H. A. Heering, J. Hirst and F. A. Armstrong, Interpreting the catalytic voltammetry of electroactive enzymes adsorbed on electrodes, J. Phys. Chem. B, 102, 6889–6902 (1998). F. A. Armstrong and G. S. Wilson, Recent developments in Faradaic bioelectrochemistry, Electrochim. Acta, 45, 2623–2645 (2000). A. Sucheta, B. A. C. Ackrell, B. Cochrane and F. A. Armstrong, Diode-like behavior of a mitochondrial electron transport enzyme, Nature, 356, 361–362 (1992). J. Hirst, J. L. C. Duff, G. N. L. Jameson et al. Kinetics and mechanism of redox-coupled long range proton transfer in an iron–sulfur protein. Investigation by fast-scan protein-film voltammetry, J. Am. Chem. Soc., 120, 7085–7004 (1998). F. A. Armstrong, Voltammetric investigation of iron–sulfur clusters in proteins, in J. Q. Chambers and A. Brajter-Toth (eds.), Electroanalytical Methods for Biological Materials. Marcel Dekker, New York, 2002, pp. 143–195. J. Hirst, B. A. C. Ackrell, F. A. Armstrong, Global observation of hydrogen-deuterium isotope effects in bidirectional electron transport in an enzyme, J. Am. Chem. Soc., 119, 7434–7439 (1997). H. A. Heering, J. H. Weiner and F. A. Armstrong, Direct detection and measurement of electron relays in a multicentered enzyme, J. Am. Chem. Soc., 119, 11628–11638 (1997). M. S. Mondal, D. G. Goodin and F. A. Armstrong, Simultaneous voltammetric comparsions of reduction potentials, reactivities and stabilities of the high potential catalytic states of wild type and distal pocket mutant (W51F) yeast cytochrome c peroxidase, J. Am. Chem. Soc., 120, 6270–6276 (1998). L. J. Anderson, D. J. Richardson and J. N. Butt, Catalytic protein film voltammetry from a respiratory reductase provides evidence for complex electrochemical modulation of enzyme activity, Biochemistry, 40, 11294–11307 (2001). J. N. Butt, L. J. Anderson, L. M. Rubio, D. J. Richardson, E. Flores and A. Herrero, Enzyme-catalyzed nitrate reduction – themes and variations as revealed by protein film voltammetry, Bioelectrochemistry, 56, 17–18 (2002). S. J. Elliot, K. R. Hoke, K. Heffron et al. Voltammetric studies of the catalytic mechanism of respiratory nitrate reducase from Escherichia coli: how nitrate reduction and inhibition depend on the oxidation state of the active site, Biochemistry, 43, 799–807 (2004). S. J. Elliot, C. Leger, H. R. Perdash et al. Detection and interpretation of redox potential optima in the catalytic activity of enzymes, Biochim. Biophys. Acta, 1555, 54–59 (2002). B. Frangioni, P. Arnoux, M. Sabaty et al. Rhodobacter spaeroides repiratory nitrate reductase, the kinetics of substrate binding favors intramolecular electron transfer, J. Am. Chem. Soc., 126, 1328–1329 (2004). H. C. Angove, J. A. Cole, D. J. Richardson and J. N. Butt, Protein film voltammetry reveals distinct fingerprints of

Electrochemistry of Redox Enzymes

195.

196.

197.

198.

199.

200.

201.

202.

203.

204.

205. 206.

207.

208.

nitrite and hyroxylamine reduction by a cytochrome c nitrite reductase, J. Biol. Chem., 277, 23374–23381 (2002). J. D. Gwyer, D. J. Richardson and J. N. Butt, Resolving complexity of interactions of redox enzymes and their inhibitors: contrasting mechanisms for the inhibition of a cyt c nitrite reductase revealed by protein film voltammetry, Biochemistry, 43, 15086–15094 (2004). K. Draoui, P. Bianco, J. Haladjian, F. Guerlesquin and M. Bruschi, Electrochemical investigation of intermoleclar electron transfer between two physiological redox partners. Cyt c3 and immobilized hydrogenase from Desulfovibrio desulfricanus Norway, J. Electroanal. Chem., 313, 201–214 (1991). C. Leger, A. K. Jones, W. Rosebloom, S. P. J. Albracht and F. A. Armstrong, Enzyme electrokinetics: hydrogen evolution and oxidation by Allochromatium vinosum [NiFe]¼hydrogenase, Biochemistry, 41, 15736–15746 (2002). K.-F. Aguey-Zinsou, P. V. Bernhardt, U. Kappler and A. G. McEwan, Direct electrochemistry of a sulfite dehydrogenase, J. Am. Chem. Soc., 125, 530–535 (2003). S. J. Elliot, A. E. McElhaney, C. Feng, J. H. Enemark and F. A. Armstrong, A voltammetric study of interdomain electron transfer within sulfite oxidase, J. Am. Chem. Soc., 124, 11612–11613 (2002). J. N. Butt, M. Filipiak and W. R. Hagen, Direct electrochemistry of M. elsdenii iron hydrogenase, Eur. J. Biochem., 254, 116–122 (1997). K. Hoke, N. Cobb, F. A. Armstrong and R. Hille, Electrochemical studies of arsenite oxidase: an unusual example of a highly co-operative two-electron molybdenum center, Biochemistry, 43, 1667–1674 (2004). K. F. Aguey-Zinsou, P. V. Bernhardt and S. Leimk€uhler, Protein film voltammetry of Rhodobacter capsulatus xanthine dehydrogenase, J. Am. Chem. Soc., 125, 15352–15358 (2003). Y. Zhou, N. Hu, Y. Zeng and J. F. Rusling, Heme proteinclay films: direct electrochemistry and electrochemical catalysis, Langmuir, 18, 211–219 (2002). C. Lei, U. Wollenberger, C. Jung and F. W. Scheller, Claybridged electron transfer between cytochrome P450cam and electrode, Biochem. Biophys. Res. Commun., 268, 740–744 (2000). E. F. Bowden, Wiring mother nature, Interface, 6, 40–44 (1997). A. El Kasmi, J. M. Wallace, E. F. Bowden, S. M. Binet and R. J. Linderman, Controlling interfacial electron-transfer kinetics of cytochrome c with mixed self-assembled monolayers, J. Am. Chem. Soc., 120, 225 (1998). J. N. Butt, J. Thornton, D. J. Richardson and P. S. Dobbin, Voltammetry of a flavocytochrome c3: the lowest potential heme modulates fumarate reduction rates, Biophys. J., 78, 994–1009 (2000). A. S. Haas, D. L. Pilloud, K. S. Reddy et al. Cytochrome c and cytochrome c oxidase: monolayer assemblies and catalysis, J. Phys. Chem. B., 105, 11351–11362 (2001).

83

209. D. D. Schlereth, Characterization of protien monolayers by surface plasmon resonance combined with cyclic voltammetry ‘in situ’, J. Electroanal. Chem., 464, 198 (1999). 210. D. L. Johnson, J. L. Thompson, S. K. Brinkmann, K. A. Schuller and L. L. Martin, Electrochemical characterization of purified Rhus vernicifera laccase: voltammetric evidence for a sequential four-electron transfer, Biochemistry, 42, 10229–10237 (2003). 211. H. A. Heering, F. G. M. Wiertz, C. Dekker and S. de Vries, Direct immobilization of native yeast iso-1 cyt c on bare gold: fast electron relay to redox enzymes and zeptomole protein-film voltammetry, J. Am. Chem. Soc., 126, 11103–11112 (2004). 212. A. Bardea, E. Katz, A. F. B€uckmann and I. Willner, NAD þ dependent enzyme electrodes: electrical contact of cofactor-dependent enzymes and electrodes, J. Am. Chem. Soc., 119, 9114–9119 (1997). 213. M. Zayats, E. Katz and I. Willner, Electrical contacting of flavoenzymes and NAD(P) þ -dpeendent enzymes by reconsititution and affinity interactions on phenylboronic monolayers assosciated with Au electrodes, J. Am. Chem. Soc., 124, 14724–14735 (2002). 214. I. Willner, V. Hegel-Shabtai, E. Katz, H. K. Rau and W. Haehnel, Integration of a reconstituted de novo synthesized hemoprotien and native metalloproteins with electrode supports for bioelectronic and bioelectrocatalytic applications, J. Am. Chem. Soc., 121, 6455–6468 (1999). 215. F. Patolsky, M. Zayats, E. Katz and I. Willner, Precipitation of an insoluble product on enzyme monolayer for biosensor applications, Anal. Chem., 71, 3171–3180 (1999). 216. L. Alfonta, E. Katz and I. Willner, Sensing of acetylcholine by a tricomponent-enzyme layered electrode using Faradaic impedance spectroscopy, cyclic voltammetry, and microgravimetric quartz crystal microbalance transduction methods, Anal. Chem., 72, 927–935 (2000). 217. Y. Xiao, F. Patolsky, J. F. Hainfeld, E. Katz and I. Willner, Plugging into enzymes: nanowiring of redox enzymes by a gold nanoparticle, Science, 299, 1877–1881 (2003). 218. E. Katz and I. Willner, Biomolecule functionalized carbon nanotubes: applications in nanobioelectronics, ChemPhysChem., 5, 1084–1104 (2004). 219. X. Yu, D. Chattopadhyay, I. Galeska, F. Papadimitrakopoulos and J. F. Rusling, Peroxidase activity of enzymes bound to the ends of single-wall carbon nanotube forest electrodes, Electrochem. Commun., 5, 408–411 (2003). 220. J. J. Davis, R. J. Coles and H. A. O. Hill, Protein electrochemistry at carbon nanotube electrodes, J. Electronanal. Chem., 440, 279–282 (1997). 221. A. Guiseppe-Elie, C. Lei and R. H. Baughman, Direct electron transfer of glucose oxidase on carbon nanotubes, Nanotechnology, 13, 559–564 (2002). 222. S. G. Wang, Q. Zhang, R. Wang et al. Multi-walled carbon nanotubes for the immobilization of enzyme in glucose biosensors, Electrochem. Commun., 5, 800–803 (2003). 223. S. Hrapovic, Y. Liu, K. B. Male and J. H. T. Luong, Electrochemical biosensing platforms using platinum nano-

84

224.

225.

226.

227.

228.

229.

230.

231.

232. 233.

234.

235.

236.

237.

238.

239.

Bioelectrochemistry

particles and carbon nanotubes, Anal. Chem., 76, 1083–1088 (2004). B. R. Azamian, J. J. Davis, K. S. Coleman, C. B. Bagshaw and M. L. H. Green, Bioelectrochemical single-wall carbon nanotubes, J. Am. Chem. Soc., 124, 12664–12665 (2002). J. Wang and M. Musameh, Carbon nanotube/teflon composite electrochemical sensors and biosensors, Anal. Chem., 75, 2075–2079 (2003). K. Yamamoto, G. Shi, T. Zhou et al. Study of carbon nanotubes-HRP modified electrode and its application for novel on-line biosensors, Analyst, 128, 249–254 (2003). R. R. Moore, C. E. Banks and R. G. Compton, Basal plane pyrolytic graphite modified electrodes: comparison of carbon nanotubes and graphite powder as electrocatalysts, Anal. Chem, 76, 2677–2682 (2004). X. Yu, D. Chattopadhyay, I. Galeska, F. Papadimitrakopoulos and J. F. Rusling, Peroxidase activity of enzymes bound to the ends of single-wall carbon nanotube forest electrodes, Electrochem. Commun., 5, 408–411 (2003). X. Yu, S. Kim, F. Papadimitrakopoulos and J. F. Rusling Protein immunosensor using single-wall carbon nanotube forests with electrochemical detection of enzyme labels, Molec. Biosys., 1, 70–78. F. Patolsky, Y. Weizmann and I. Willner, Long-range electrical contacting of redox enzymes by SWCNT connectors, Angew. Chem. Int. Ed., 43, 2113–2117 (2004). J. T. Elliot, C. W. Meuse, V. Silin et al. Biomimetic membranes on metal supports, in J. F. Rusling (ed.), Biomolecular Films. Marcel Dekker, New York, 2003, pp. 99–162. J. Israelachvili, Intermolecular and Surface Forces, 2nd edn., Academic Press, San Diego, 1992. H. T. Tien and Z. Salamon, Formation of self-assembled lipid bilayers on solid substrates, Bioelectrochem. Bioenerg., 22, 211–218 (1989). T. Martynski and H. T. Tien, Spontaneous assembly of bilayer membranes on a solid surface, Bioelectrochem. Bioenerg., 25, 317–342 (1991). A. Guerrieri, T. R. I. Cataldi and H. A. O. Hill, Direct electrical communication of cytochrome c and cytochrome b5 at basal plane graphite electrodes modified with lauric acid or laurylamine, J. Electroanal. Chem., 297, 541–547 (1991). Z. Salamon, F. K. Gleason and G. Tollin, Direct electrochemistry of thioredoxins and glutathione at a lipid bilayer modified electrode, Arch. Biochem. Biophys., 299, 193–198 (1992). Z. Salamon and G. Tollin, Cyclic voltammetric behavior of [2Fe-2S] ferredoxins at a lipid bilayer modified electrode, Bioelectrochem. Bioenerg., 27, 381–391 (1992). Z. Salamon and G. Tollin, Interaction of horse heart cytochrome c with lipid bilayer membranes: effects on redox potentials, J. Bioenerg. Biomembranes, 29, 211–221 (1997). Z. Salamon, J. T. Hazzard and G. Tollin, Direct measurement of cyclic current-voltage respneses of integral membrane proteins at a self-assembled lipid bilayer modified

240.

241.

242.

243.

244.

245.

246.

247.

248. 249.

250.

251.

252.

253.

electrode: cytochrome f and and cytochrome c oxidase, Proc. Natl. Acad. Sci. USA, 90, 6420–6423 (1993). Z. Salamon, G. Tollin, M. Hirisawa et al. The oxidationreduction properties of spinach thioredoxins f and m and of ferredoxin:thioredoxin reductase, Biochim. Biophys. Acta, 1230, 114–118 (1995). K. T. Kinnear and H. G. Monbouquette, Direct electron transfer to E. coli fumarate reductase in self-assembled alkanethiol monolayers on gold electrodes, Langmuir, 9, 2255–2257 (2000). K. T. Kinnear and H. G. Monbouquette, An amperometric fructose biosensor based on fructose dehydrogenase immobilized in a membrane mimetic layer on gold, Anal. Chem., 69, 1771–1775 (1997). J. K. Cullison, F. M. Hawkridge, N. Nakashima and S. Yoshikawa, A study of cytochrome c oxidase in lipid bilayer films on electrode surfaces, Langmuir, 10, 877–882 (1994). J. D. Burgess and F. M. Hawkridge, Octadecyl mercaptan sub-monolayers on silver electrodeposited on gold quartz crystal microbalance electrodes, Langmuir, 13, 3781–3786 (1997). J. D. Burgess, M. C. Rhoten and F. M. Hawkridge, Cytochrome c oxidase immobilized in stable supported lipid bilayer membranes, Langmuir, 14, 2467–2475 (1998). J. D. Burgess, M. C. Rhoten and F. M. Hawkridge, Observation of the resting and pulsed states of cytochrome c oxidase in electrode- supported lipid bilayer membranes, J. Am. Chem. Soc., 120, 4488–4491 (1998). R. Santucci, T. Ferri, L. Morpurgo, I. Savani and L. Avigliano, Unmediated heterogenous electron transfer of ascorbate oxidase and laccase at a gold electrode, Biochem. J., 332, 611–615 (1998). J. F. Rusling, Enzyme bioelectrochemistry in cast biomembrane-like films, Acc. Chem. Res., 31, 363–369 (1998). A.-E. F. Nassar, Z. Zhang, J. F. Rusling et al. Orientation of myoglobin in cast multibilayer membranes of amphiphilic molecules, J. Phys. Chem., 99, 11013–11017 (1995). X. Zu, Z. Lu, Z. Zhang, J. B. Schenkman and J. F. Rusling, Electro-enzyme catalyzed oxidation of styrene and cis-bmethylstyrene using thin films of cytochrome P450cam and Myoglobin, Langmuir, 15, 7372–7377 (1999). E. I. Iwuoha, S. Joseph, Z. Zhang et al. Drug metabolism biosensors 3: electrochemical reactivities of cyt P450cam immobilized in synthetic vesicular systems, J. Pharm. Biomed. Anal., 17, 1101–1110 (1998). E. I. Iwuoha, S. Joseph, Z. Zhang and M. Smyth, Reactivities of organic phase biosensors 6: square wave and differential pulse studies of genetically engineered cyt P450cam bioelectrodes in selected solvents, Biosens. Bioelectron., 18, 237–244 (2003). C. E. Immoos, J. Chou, M. Bayachou et al. Electrocatalytic reductions of nitrite, nitric oxide, by thermophilic cyt P450 CYP119 in film-modified electrodes and an analytical comparison of its catalytic activities with myoglobin, J. Am. Chem. Soc., 126, 4934–4942 (2004).

Electrochemistry of Redox Enzymes

254. E. Blair, J. Greaves and P. J. Farmer, High-temperature electrocatalysis using thermophilic cyt P450 CYP119: dehalogenation of CCl4 to CH4, J. Am. Chem. Soc., 126, 8632–8633 (2004). 255. M. Bayachou and J. A. Boutros, Direct electron transfer to the oxygenase domain of neuronal nitric oxide synthetase (NOS): exploring unique redox properties of NOS enzymes, J. Am. Chem. Soc., 126, 12722–12723 (2004). 256. Z. Zhang, S. Chouchane, R. S. Magliozzo and J. F. Rusling, Direct voltammetry and enzyme catalysis with M. tuberculosis catalase-peroxidase, peroxidases and catalase in lipid films, Anal. Chem., 74, 163–170 (2002). 257. K. F. Aguey-Zinsou, P. V. Bernhardt, A. G. McEwan and J. P. Ridge, The first non-turnover voltammetric response from a molybenum enzyme: direct electrochemistry of dimethylsulfoxide reductase from Rhodobacter capsulatus, J. Biol. Inorg. Chem., 7, 879–883 (2003). 258. P. V. Bernhardt, G. Schenk and G. J. Wilson, Direct electrochemistry of procine purple acid phosphatase (Uteroferrin), Biochemistry, 43, 10387–10392 (2004). 259. B. Munge, Z. Pendon, H. A. Frank and J. F. Rusling, Electrochemical reactions of redox cofactors in Rhodobacter sphaeroides reaction center proteins in lipid films, Bioelectrochem., 54, 145–150 (2001). 260. Z. Gao, H. A. Frank, Y. M. Lvov and J. F. Rusling, Influence of bromide on electrochemistry of photosynthetic reaction center films on gold electrodes, Bioelectrochem., 54, 97–100 (2001). 261. P. Bianco and J. Haladjian, Electrocatalytic hydrogen evolution at the pyrolytic graphite electrode in the presence of hydrogenase, J. Electrochem. Soc., 139, 2428–2432 (1992). 262. J. F. Rusling, Electroactive and enzyme-active proteinpolyion films assembled layer-by-layer, in Y. Lvov and H. M€ ohwald (eds.), Protein Architecture: Interfacing Molecular Assemblies and Immobilization Biotechnology. Marcel Dekker, New York, 2000, pp. 337–354.

85

263. Y. M. Lvov, Z. Lu, J. B. Schenkman, X. Zu and J. F. Rusling, Direct electrochemistry of myoglobin and cytochrome P450cam in alternate polyion layer-by-layer films with DNA and other polyions, J. Am. Chem. Soc., 120, 4073–4080 (1998). 264. H. Ma, N. Hu and J. F. Rusling, Electroactive myoglobin films grown layer-by-layer with poly(styrenesulfonate) on pyrolytic graphite electrodes, Langmuir, 16, 4969–4975 (2000). 265. Z. Lu, Y. Lvov, I. Jansson, J. B. Schenkman and J. F. Rusling, Electroactive films of alternately layered polycations and iron–sulfur protein putidaredoxin on gold, J. Coll. Interface Sci., 224, 162–168 (2000). 266. P. He, N. Hu and J. F. Rusling, Driving forces for layer-bylayer self assembly of films of SiO2 nanoparticles and heme proteins, Langmuir, 20, 722–729 (2004). 267. A. C. Onuoha, X. Zu and J. F. Rusling, Electrochemical generation and reactions of ferrylmyoglobin in water and microemulsions, J. Am. Chem. Soc., 119, 3979. (1997). 268. B. Munge, C. Estavillo, J. B. Schenkman and J. F. Rusling, Optimizing electrochemical and peroxide-driven oxidation of styrene with ultrathin polyion films containing cytochrome P450cam and myoglobin, Chem. Biochem., 4, 82–89 (2003). 269. S. Joseph, J. F. Rusling, Y. M. Lvov, T. Friedberg and U. Fuhr, An amperometric biosensor with immobilized human cytochrome P450 (CYP3A4) as a novel drug screening tool, Biochemical Pharmacology, 65, 1817–1826 (2003). 270. C. Estavillo, Z. Lu, I. Jansson, J. B. Schenkman and J. F. Rusling, Epoxidation of styrene by human cyt P450 1A2 using thin film electrolysis compared to conventional solution reactions, Biophys. Chem., 104, 291–296 (2003). 271. For examples, see J. F. Rusling and T. F. Kumosinski, Nonlinear Computer Modeling of Chemical and Biochemical Data, Academic Press, San Diego, 1996, pp. 196–205.

3

Biological Membranes and Membrane Mimics Tibor Hianik Department of Nuclear Physics and Biophysics, Comenius University, Bratislava, Slovakia

3.1 INTRODUCTION The biological membrane is one of the most important cell structures. It represents an envelope for the cell with a unique barrier function, that provides directional transport of species into the cell, and waste and toxic compounds out of the cell. In addition, the low permeability of the membrane for charged particles, e.g. ions, allows a non-equilibrium ion distribution to be main between the extra cellular and the cytoplasmic sides of the cell, which is crucial for cell function. Destruction of the membrane results in the establishment of equilibrium and cell apoptosis. The membrane, with a supported protein net – glycocalix is responsible for the cell shape and, owing to its viscoelasticity, also for reversible changes of shape during cell function. However biomembranes provide not only structural and barrier functions. They contain integral and peripheral proteins that are responsible for communication with the surrounding environment. They provide receptor functions and are also responsible for the transfer of the signals into the cell by means of sophisticated signaling pathways. Several catalytic processes are also concentrated in the membrane, for example the energy transduction connected with synthesis of the energetically rich molecule adenosine triphosphate (ATP). From a physical point of view, the biomembrane represents an anisotropic and inhomogeneous structure, with properties typical of smectic-type liquid crystals. Due to the rather complicated structure, anisotropy and inhomogeneity, the study of the physical and electrochemical properties of

biomembranes is difficult. Therefore, several models of biomembranes have been developed, including micelles, monolayers, lipid bilayers, liposomes and also solid supported lipid films. These structures allow the large-scale variation of the lipid composition and the incorporation of integral or peripheral proteins. Thus, model membranes can be constructed in a way that mimics the structure and properties of biomembranes. Over the last few decades, the unique properties of lipid membranes have allowed the fabrication of biocompatible and biofunctional interfaces on solid surfaces. These supported lipid films allow the immobilization of various functional macromolecules, like enzymes, antibodies, receptors and nucleic acids, without the loss of their native conformation, selectivity, sensitivity and catalytic activity. These systems allow stress-free analysis of the interactions of various pharmacological drugs with the membrane and thus the effect of these compounds on cell behavior. The lipid films are selfassembling structures. This unique property can be utilized in the fabrication of smart biosensors with excellent sensitivity and selectivity. This chapter is devoted to introducing the peculiarities of these exciting structures and to demonstrating the unique physical and electrochemical properties of biomembranes and their models. The reader is introduced step by step to the features of the membrane structure and historically, how this structure has been, established. New knowledge obtained in recent studies is presented. The chemical composition of biomembranes is then shown. Model structures, allowing the study of physical properties

Bioelectrochemistry: Fundamentals, Experimental Techniques and Applications Edited by Philip Bartlett # 2008 John Wiley & Sons, Ltd. ISBN: 978-0-470-84364-2

88

Bioelectrochemistry

of membranes, such as monolayers, bilayers, liposomes and supported membranes, are described together with typical methods used for their study. Important phenomena and properties, such as electrical and mechanical stability, electroporation, membrane thermodynamics and mechanics, protein–lipid interactions, membrane potentials, ionic transport, cell receptors and signaling are considered. Examples of the applications of supported lipid membranes in fundamental bioelectrochemistry, as well as in applied research directed at construction of biosensors, are also presented. 3.2 MEMBRANE STRUCTURE AND COMPOSITION 3.2.1 Membrane Structure Biological membranes are one of the most important structural and functional components of the cell. They fulfill a number of important functions [1–4]: (1) Structural – they surround the cell cytoplasm and give a certain form to the cell and its organelles. (2) Barrier – they secure the passage into and out of the cell of only necessary ions, low molecular compounds, proteins etc. (3) Contact – they facilitate contact between cells by means of specific structures. (4) Receptors – they are susceptible to different signals from the surrounding environment by means of special protein structures incorporated into the membrane. These signals can be light, mechanical deformations, specific substances etc. (5) Transport – they provide the active and passive transmembrane membrane transport of ions, glucose, amino acids and other compounds, as well as transport of electrons in mitochondria and chloroplasts. The structures of all biomembranes have a whole series of common features (Figure 3.1). Their basis is a lipid bilayer composed of lipid molecules into which are incorporated peripheral and integral proteins. They are supported or covered by structural proteins, such as the spectrin net in erythrocytes [2], or bacterial S-proteins [5]. The outer parts of bacterial and plant membranes are also covered by polysaccharides [2]. The lipid matrix provides the integrity of the membrane, electrical isolation, and the possibility of self-assembly of the corresponding protein structure in the membrane. The proteins determine the fulfillment of the specific functions by the membrane. In particular, the integral proteins, which penetrate through the membrane, are, for

Figure 3.1 A model of the structure of a biomembrane.

example, ionic channels or the proteins that provide active ionic transport – ATPases, etc. In normal, physiological conditions, the lipid bilayer is in a liquid-crystalline state. From a physical point of view, the biomembrane represents a smectic-type liquid crystal [6]. The thickness of biomembranes varies from 5–10 nm and is considerably less than the dimension of the cells. The variations in biomembrane thickness are mostly due to the integral, peripheral and structural proteins, as well as the presence of lipopolysaccharides and glycopolysaccharides [2]. Due to the thinness of these membranes, the establishment of their existence and the subsequent study of their structure and properties were not easy. Progress in understanding the peculiarities of the membrane structure and properties was directly connected with progress in physics. The appearance of the light microscope and especially the considerable progress in the fabrication of optical lenses in the 17th century allowed the first observation of the cell structure by Robert Hooke in 1662. However, further progress was rather slow and it wasn’t until 1831 that Robert Brown showed the existence of a nucleus in the cell [7]. Then cell theory, one of the fundamental concepts of biology, was formulated by botanist (Schleiden, 1838) and zoologist (Schwann, 1839) [2]. However, at this time even a hypothesis of the existence of the membrane did not exist. The existence of the membrane surrounding the cell was only proposed in 1855 by Negeli, who observed that undamaged cells can change volume upon changes in the osmotic pressure of the surrounding environment. These experiments were continued by Overton [1], who showed that non-polar molecules penetrate

Biological Membranes and Membrane Mimics

89

Table 3.1 The composition of biomembranes [1]. % of dry weight

Biomembrane Myelin Human erythrocytes Outer segment of retina rood Mitochondria of the rat liver Inner membranes Outer membranes Salmonella typhimurium Inner membranes Outer membranes

Figure 3.2 A brief schematic historical overview of the models of biomembranes: (a) lipid bilayer according to Gortel and Grendel [8]; (b) model by Danielli and Dawson [11]; (c) unitary model of biomembrane by Robetson [14,15]; (d) micellar model proposed by Lucy [12].

more easily through the cell membrane than polar molecules. On the basis of these experiments, he raised the hypothesis that the membrane structure had a lipid nature. Further development of the concept of membrane structure was achieved thanks to work by Gorter and Grendel [8] in the first third of the 20th century. This time was characterized by excellent achievements in the study of monomolecular layers at air–water interfaces mostly due to the work by Langmuir and coworkers [9,10]. Gorter and Grendel [8] used this approach. They extracted lipids from erythrocytes and showed that the area of the monomolecular layer formed by lipids at an air–water interface is twice the area of the erythrocyte cell. This resulted in the concept that the biomembrane is composed of two monomolecular lipid layers (Figure 3.2a). Despite certain errors in their study, which consisted in underestimation of the concentration of the lipids (they used acetone for extraction of lipids, which does not allow the extraction of all the lipids) and the area of erythrocytes (they determined the area from dry cells), as well as not considering the presence of proteins, the bilayer concept of biomembrane structure was accepted by the wider scientific community. It is now clear that with more accurate experiments, the interpretation would be different. It is known now that the membranes of erythrocytes are composed of only 43 % lipids, while the rest, 49 %, is protein and 8 % hydrocarbons (see Table 3.1). The assumption that proteins are also connected to membranes was raised 10 years later by Danielli and

proteins

lipids

hydrocarbons

18 49 51 – 76 52 – 65 44

79 43 49 – 24 48 – 35 20

3 8 – – 2< – – – 37

coworkers [11] due to the necessity for explaining the substantially lower surface tension of biomembranes in comparison with pure lipid monolayers at air–water interfaces. For example, the surface tension of the membrane of sea urchin cells was approx. 0.2 mN m1, while that for monolayers of fatty acids at the air–water interfaces was 10–15 mN m1 [12]. Further studies also showed that addition of proteins to the water sub-phase resulted in a decrease in the surface tension of lipid monolayers. It was therefore proposed that globular proteins are connected with both surfaces of the lipid bilayer (Figure 3.2b). Direct evidence of the existence of biomembranes was possible only after the discovery of the electron microscopy and its application in biology in the 1950s. The first electron micrographs showed that the cell is surrounded by a thin membrane approx. 6–10 nm in thickness. This membrane is composed of three layers. Two high electron density layers approx. 2 nm thick are separated by a low electron density layer of thickness 3.5 nm [13]. A similar structures was also observed in most intracellular organelles [14]. It was proposed that the high electron density layers correspond to the region of polar head groups of the phospholipids covered by proteins, while the low electron density layer corresponds to the hydrophobic part of the lipid bilayer (see [7] for electron micrographs of cell structures). On the basis of these results, J. Robertson [15] proposed the elementary model of the asymmetric cell membrane. According to this model, the lipid bilayer is covered by a layer of proteins in a b-conformation that are adjacent to the polar part of the membrane due to electrostatic interactions. The asymmetry is due to the fact that the outer monolayer is covered by glycoproteins (Figure 3.2c). Further, Lucy [12] proposed the membrane model composed of micelles covered by proteins (Figure 3.2d). However, this model cannot explain the

90

Bioelectrochemistry

rather small conductivity of bilayer lipid membranes (BLM) determined by Mueller et al. [16]. This and further studies showed that the specific conductance of the BLM is in the range 106–1010 S cm2 [17,18]. High resolution electron microscopy as well as improved methods of preparing ultra thin samples allowed micrographs of cell membranes with additional structural details to be obtained. The existence of channels in a membrane as well as the mosaic structure of the cell surface was shown. The analysis of these results led to the proposal of the so-called fluid-mosaic model of biomembranes [19,20]. This model, shown in Figure 3.1, considers the membrane to be a continuous lipid bilayer with incorporated proteins. The model explains in particular, the dependence of the activity of membrane proteins on the physical state of the membrane as well as the existence of the membrane viscosity. Further, the proteincrystallic model proposed by Vanderkooi and Green [21] differs from the fluid-mosaic model only by postulating the existence of a rigid protein structure in the membrane. Currently various variations of fluid- mosaic models are used to describe biomembrane structure. However, studies of the physical properties of biomembranes and lipid bilayers have revealed that the mobility of some membrane proteins is strongly restricted. Also the concept of continuous a lipid bilayer is a certain simplification. Experimental and theoretical studies performed recently have shown that the lateral structure of a lipid bilayer is dynamic and heterogeneous, and is characterized by a lipid-domain formation with different mobilities for lipids and proteins [1]. Specific interactions between membrane components lead to selective orientation and segregation of the lipids and proteins in the plane of the membrane. There are lipid clusters composed of up to several hundreds molecules. The existence of long-range superstructure was proved on model membrane systems by several methods, e.g. by scanning tunneling microscopy [22]. Aggregates of proteins are also an important peculiarity of biomembrane structure. A typical example of this phenomenon is the aggregation of an integral protein, bacteriorhodopsin, in membranes [23]. The two-dimensional matrix of a biomembrane probably consists of patches of phospholipid molecules in different degrees of conformational disorder. Under certain conditions, the bilayer organization can be interrupted by non-bilayer phases [24], as well as by bilayer phases of different composition, and by regions of mismatch between coexisting phases. Such features within the organization of a membrane can have different lifetimes, and may be induced in response to environmental and metabolic perturbation [4]. Several types of

molecular motion of lipids and proteins are experienced by the components within the membrane: rotation of molecules along their axes perpendicular to the plane of the membrane occurs every 0.1–100 ns for lipids and 0.01–100 ms for proteins; segmental motion of acyl chains (0.01–1 ns) gives rise to an increased disorder toward the center of the membrane; translational motion of molecules in the plane of the membrane occurs with a lateral diffusion coefficient of 1013 to 108 cm2 s1. These orientational and motional parameters for the components in the membrane differ more than what would be expected only on the basis of the size of these components. It should be emphasized that not all the molecules of the same type in the same membrane necessarily have the same motional properties [4]. Biomembranes are composed of lipids, proteins and hydrocarbons. The proteins and lipids represent the main part of the biomembrane. The content of hydrocarbons usually does not surpass 10 %. Hydrocarbons are mostly covalently bonded to the lipids (glycolipids) or to the proteins (glycoproteins). Both glycolipids and glycoproteins play an important role in cell recognition. The hydrocarbons are localized in the outer part of all biomembranes and thus, together with the different chemical composition of lipids at both membrane monolayers, contribute to the membrane asymmetry. The membrane is also asymmetrical with respect to proteins. The content of proteins in the membrane varies from almost 20 % for myelin membranes, to almost 80 % for inner membranes of mitochondria (Table 3.1). 3.2.2 Membrane Lipids A lipid bilayer represents a self-assembled structure formed from lipids in an aqueous solution. This is the result of the hydrophobic effect, whereby the non-polar acyl chains of lipids (which form the interior of the bilayer) and the non-polar amino acid residues in proteins tend to be squeezed away from the aqueous phase. There exist more then 100 phospholipids that are differ in their polar head groups and the composition of their hydrophobic chains. Lipids can be divided into three main classes: phospholipids, glycolipids and sterols. Phospholipids are the most common lipids in cell membranes. They are divided into two main classes: glycerophospholipids and sphingophospholipids (derivatives of ceramide and sphingomyelin). The glycerophospholipids (phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylserine (PS), phosphatidylinostitol (PI) and phosphatidylglycerol (PG)) have a similar construction, which consists of a polar head group and two hydrophobic chains of fatty acids that are connected to the glycerol

Biological Membranes and Membrane Mimics

91

Figure 3.3 The structure of a glycerophospholipid, phosphatidylcholine: (a) chemical structure; (b) schematic representation. The glycerophospholipid is composed of five parts: a hydrophilic choline head group is connected through the phosphate residue to the glycerol backbone. The glycerol is connected to two hydrocarbon chains of fatty acids. This part creates the hydrophobic part of the phospholipid. At the double bond the hydrophobic chain is tilted in respect to the direction normal to the membrane plane.

backbone (Figure 3.3). The main representatives of phospholipids are shown in Figure 3.4. The structures of glycerophospholipids and sphingophospholipids differ considerably in the interfacial and hydrophobic parts (compare structure of PC and SM in Figure 3.4). The most common base in mammalian SM is sphingosine (1,3-dihydroxy-2amino-4-octadecene), with a trans double bond between the C4 and C5 atoms. Phospholipids play a dominant structural role in the membrane, providing a barrier to passive translocation of ions and other species through the membrane, and providing a favourable environment for functioning membrane proteins. However, certain lipids can also play a functional role. Typical examples are phosphatidylinositol and sphingomyelin. Phosphatidylinositol is localized in the cytoplasmic side of the membrane and is important for cell signaling. Sphingomyelin is an important component of the eukaryotic cell. SM has a cylindrical shape like PC, which helps to minimize free energy in the formation of lipid bilayers. However, in addition to the structural role, it also participates in cell signaling (see Section 3.10.2.). Products of SM metabolism, like ceramide sphingosine, sphingosine-1-phosphate and diacylglycerol, are important cellular effectors and give SM a role in cellular functions like apoptosis, ageing

Figure 3.4 Structural formulae of the main lipids presented in biomembranes. PC – phosphatidylcholine, PE – phosphatidylethanolamine, PS – phosphatidylserine, PI – phosphatidylinositol, PG – phosphatidylglycerol, SM – sphingomyelin, MG – monogalactosyldiglyceride, GC – galactosylceramide, DPG – diphosphatidylglycerol (cardiolipin).

and development [25]. SM forms more stable complexes with cholesterol in comparison with other phospholipids. Results obtained during the last decade show a substantial lateral organization of both lipids and proteins in biomembranes. Sphingolipids, including SM, together with cholesterol,havebeenshowntobeimportantfactorsintheformation of lateral domains or ‘rafts’ in biological membranes. These

92

Bioelectrochemistry

domainshavebeensuggested totakepartincellular processes, such as signal transduction, membrane tracking and protein sorting. The formation of lateral ‘rafts’ in biological membranes is supposed to be driven by lipid–lipid interactions, which are largely dependent on the structure and biophysical properties of the lipid components [25,26]. Glycolipids are localized mostly in plasmatic membranes and exclusively at the extracellular side. The sugar residues of glycolipids are therefore exposed to the external part of the cell and create the protective film which surrounds most living cells. Glycolipids are represented by cerebrosides, sulphatides and gangliosides. As an example, monogalactosyldiglyceride (MG) and galactosylceramide (GC) are shown in Figure 3.4. Sterols in membranes are constructed on a sterol backbone. Among the sterols cholesterol (CH) is only typical for animal cells; it is not found in bacterial and plant cells, which contain ergosterol (ES) and stigmasterol, respectively, (Figure 3.5). The lipid composition of various living cells is presented in Table 3.2. We can see that the basic lipids are phosphatidylcholine and phosphatidylethanolamine. Glycolipids occur to a larger extent in a myelin membrane. There is a high variety of fatty acids in phospholipids. However, only two or three types of fatty acids are dominant in cell membranes. In higher plants, mostly palmitic, oleyl and linoleyl acids are found; there is practically no stearoyl acid found. The cells of living organisms also contain, in addition to palmitic and oleyl acids, fatty acids with larger numbers of carbons – 20 and more. As a rule they are composed of even numbers of

Figure 3.5 Structural formulae of the main sterols found in biomembranes. CH – cholesterol, ST – stigmasterol, ES – ergosterol.

Table 3.2 Lipid composition of cells (% of the total mass of all lipids) Lipids

Plasma membrane

Nucleus

Mitochondria

Myelin

Erythrocytes

E. Coli

Phosphatidylcholine Sphingomyelin Phosphatidylethanolamine Phosphatidylserine Phosphatidylinositol Lisophosphatidylcholine Phosphatidylglycerol Diphosphatidylglycerol Other phospholipids Cholesterol Cholesterol esters Fatty acids Glycolipids Other lipids

18.5 12.0 11.5 7.0 3.0 2.5 – 0 2.5 19.5 2.5 6.0 – 15.0

44.0 3.0 16.5 3.5 6.0 1.0 – 1.0 – 10.0 1.0 9.0 – 5.5

37.5 0 28.5 0 2.5 0 – 14.0 – – 2.5 – – 15.0

10 8.5 20.0 8.5 1.0 – – 0 – 26.0 – – 26.0 0.5

19.0 17.5 18.0 8.5 1.0 – – 0 – 25.0 – – 10.0 1.5

0 0 65.0 0 0 – 18.0 12.0 – 0 – – 0 –

Biological Membranes and Membrane Mimics

93

Table 3.3 Fatty acid composition of the phospholipids of human erythrocyte membrane. Fatty acid

PC

PE

PS

SM

C C C C C C C

34 13 22 18 6 – –

29 9 22 6 18 – –

14 36 15 7 21 – –

28 7 6 2 8 20 14

16:0 18:0 18:1 18:2 20:4 24:0 24:1

PC – phosphatidylcholines, PE – phosphatidylethanolamines, PS – phosphatidylserines, SM – sphingomyelins.

carbon atoms. The unsaturated fatty acids contain double bonds almost exclusively in a cis conformation. An example of the fatty acid composition of the membrane of human erythrocytes is presented in Table 3.3. For biomembranes phospholipids that contain unsaturated fatty acids are typical. The appearance of only one double bond in one of the fatty acid chains considerably lowers the phase transition temperature of phospholipids from a gel to a liquid-crystalline state. For example, dipalmitoylphosphatidylcholine (DPPC), composed of two saturated palmitic acids (16 carbons), has a main phase transition temperature of approx. 41  C. However, palmitoyloleylphosphatidylcholine (POPC) which differs from DPPC only by one double bond in one of the fatty acid chains has a phase transition temperature of 5  C. Thus, thanks to the unsaturated fatty acids, at physiological temperatures, biomembranes are in a liquid-crystalline state. In addition to the above mentioned lipids, there are other lipids that occur less frequently in membranes. Among these lipids are plasmalogens. In a molecule of a plasmalogen, instead of an acyl group at the first carbon atom of glycerol, there is an aldehyde group. Another phospholipid, cardiolipin (diphosphatidylglycerol) (Figure 3.4), is an important component of the membranes of mitochondria. It has been found that in cyanobacteria the nitrogen is replaced by sulfate, thus creating sulfophospholipids. These organisms are able to produce sulfocholine from cysteine and methionine, which protects cyanobacteria so they cannot be utilized by other organisms. Thermophilic and methane producing bacteria contain diphytanoyl glycerolethers in their membranes. The fatty acid chains are covalently connected in the middle of the membrane. This results in high stability of the lipid bilayer and protects the membrane from disruption at higher temperatures, as well as against the dissolution effect of methanol.

Figure 3.6 Transbilayer distribution of phospholipids in the human erythrocyte membrane. TPL – total phospholipid, SM – sphingomyelin, PC – phosphatidylcholine, PE - phosphatidyletanolamine and PS – phosphatidylserine. (According to [27]).

The distribution of the lipids in a membrane is highly asymmetric. The glycolipids are exclusively located in the outer monolayer of the membrane. The assymmetry in the distribution of phospholipids in the membrane of erythrocytes is shown in Figure 3.6. The majority of two cholinecontaining phospholipids, sphingomyelin and phosphatidylcholine is localized in the outer monolayer. Two aminophospholipids are predominantly (phosphatidylethanolamine) or even exclusively (phosphatidylserine) localized in the cytoplasmic half of the bilayer [27]. The uncatalyzed exchange of lipid molecules is very slow and probably does not exist for proteins. Transbilayer movement is energetically unfavorable because it requires the insertion of the polar groups into the nonpolar region and the exposure of the apolar groups to the polar region. Slow uncatalyzed transbilayer movement (so-called flipflop) of some phospholipids has a halflife value of 3 to 27 h (phosphatidylcholine), whereas certain phospholipids (phosphatidylethanolamine) are subject to an ATPdependent ‘flippase’ catalyzed inward movement with a half life of approximately 30 min. The catalyzed transmembrane movement of phospholipids has been found in erythorcutes, rat liver microsomes and cultured fibroblasts (see [28] and references therein), the transbilayer movement of cholesterol is probably much faster than that of any other lipid, with a half life values in the order of seconds.

94

Bioelectrochemistry

All membrane components are recycled many times during the life of the cell. The lifetime of the phospholipids depends on the intensity of function of the membrane. For example, the lifetime of phosphatidylcholine in a myelin membrane is 2 months, while in a mitochondria membrane where extensive oxidative processes take place, the lifetime is only 2 weeks.

3.2.3 Membrane Proteins Membrane proteins play an important functional role in a cell. They form ionic channels, transporters, receptors and enzymes. Certain proteins also play structural roles. An example is the spectrin net located on the cytoplasmic side of the membrane (Figure 3.7) [2]. Depending on the localization in the membrane and their role, the membrane proteins are divided into three main groups: peripheral proteins, integral proteins (see Figure 3.1) and structural proteins. Enzymes are the most widespread proteins in membranes. They can be both integral (ATPases) and peripheral (acetylcholinesterase, phosphatases). Receptors, as well as immunoproteins, can also be peripheral or integral. The receptor proteins are usually connected to additional proteins on the cytoplasmic side of the membrane to transfer the signal inside the cell. Among these proteins the G-proteins play an important role in cell signaling. Cell signaling and the role of G-proteins will be described in more detail below (see Section 3.10.2). Peripheral proteins, e.g. cytochrome c, are localized at the membrane surface and are connected to the membrane either by electrostatic interactions, or they contain a short hydrophobic chains, which anchors the peripheral protein to the membrane. Peripheral proteins can be isolated by changes in the pH or ionic strength.

Figure 3.7 Membrane proteins and their functions. A – ionic channel, B – Diffusion transporter, C – receptor, D – enzyme (transforms substrate S to product P).

Integral proteins, e.g. glycophorin, Naþ/K-ATPase or bacteriorhodopsin, are translocated across the lipid bilayer. In addition to a hydrophilic part, that contacts the water environment, they are also characterized by a hydrophobic part, that contacts the hydrophobic interior of the membrane. Integral proteins have various degrees of complexity and can pass through the membrane either once (glycophorin), or several times (bacteriorhodopsin – this protein can cross the membrane seven times [2]). Integral proteins are more tightly connected to the membrane than peripheral proteins. The membrane architecture, as well as the functioning of membrane proteins, is determined by protein–lipid interactions. This question will be considered below (Section 3.6.3). The isolation of integral proteins is more difficult compared to peripheral proteins. For this purpose it is necessary to use organic solvents or detergents. As we have already mentioned, organic solvents, such as a mixture of chloroform/ methanol, can be used to isolate integral proteins, however, after isolation the proteins are, as a rule, inactive. They can be used for structural studies, but not for functional studies. Most common is isolation of integral proteins by detergents. Among detergents, ionic detergents, like sodium dodecylsulphate and sodium cholate, or non-ionic detergents, like Triton X-100, are most common. Application of sodium cholate and Triton X-100 is, however, the mildest for preserving the function of proteins. The role of detergents consists in disturbing the lipid bilayer and in formation of detergent–protein complexes, which are soluble in water. Lipid molecules also form complexes with detergents. The molecules of detergent are conical in shape, and therefore they form micelles in water. The process of solubilization of integral proteins is showed in Figure 3.8. The disadvantage of the applicantion of detergents, however, consists in the fact that the detergents remain adjacent to the proteins, therefore additional methods of purification should be used to obtain a pure protein fraction. Structural proteins form the membrane cytoskeleton. An example is the spectrin localized at the cytoplasmic side of the erythrocyte membranes. The structural proteins are not strictly membrane proteins, but are connected to the membrane through integral proteins. For example, the band III protein of erythrocytes is connected by the small protein ankerin. The spectrin threads are connected to the ankerin (Figure 3.9). The spectrin net, together with microtubules and microfilaments, protects the cell against changes in shape or volume. The main protein of the cytoskeleton is tubulin. Tubulin is able to form aggregates and forms tube-like structures. The integrity of these structures is possible

Biological Membranes and Membrane Mimics

95

Figure 3.8 Scheme for the isolation of integral proteins by detergent.

only at very low concentrations of Ca2þ ions (usually 107 M). An increase in the calcium concentration on the cytoplasmic side of the cell can destroy the cytoskeleton. Therefore Ca2þ-pumps continuously maintain a very low level of calcium in the cell by removing it either to the cytoplasmic reticulum or outside the cell. The cytoskeleton considerably stabilizes the integrity of the cell membrane. An important factor in this stabilization is inter cellular contacts, which are created by collagen. All living cells except erythrocytes and lymphocytes have a cell envelope – the glycokalix. The lifetime of membrane proteins is from 2 to 5 days. Therefore mechanisms exist to provide transport of newly synthesized membrane proteins to the membrane. The synthesis of the proteins is started at ribosomes inside the endoplasmatic reticulum (Figure 3.10). The growth of the polypeptide chain starts from the N-terminal. First, the unique sequence of the chain recognized by a membrane receptor is synthesized. As soon as the polypeptide chain is sufficiently long, it separates from the receptor, but preserves the connection with the ribosome. After the synthesis the protein is separated from the ribosome.

3.3 MODELS OF MEMBRANE STRUCTURE The study of the physical properties of biomembranes is rather difficult due to the small size of the cell (typically

Figure 3.10 Incorporation of synthesized proteins into the membrane. A – signal peptide, B – signal peptide connected with receptor, C – receptor facilitates the incorporation of the peptide into the membrane, D – after incorporation of the peptide, the signal part is removed and polysaccharides are attached, E – ribosome is separated from the membrane [29].

several mm), the small thickness of the membrane (5–10 nm), considerable inhomogeneity and anisotropy. In addition, it is difficult to study separately the properties of the lipid bilayer and influence of proteins on the bilayer. Therefore, biophysical studies of membrane properties have been performed on various models of membrane structure, such as lipid monolayers, multilayers, bilayer lipid membranes (BLM), multi- or unilamellar vesicles and supported bilayer lipid membranes (sBLM). Historically the first models of membrane structure were lipid monolayers, which had a considerable role in establishing the bilayer nature of biomembranes in the first third of the 20th century. Stable bilayer lipid membranes were reported in 1962 by Mueller and coworkers [30]. Finally, in 1965, Bangham with coworkers [31] discovered liposomes, which became the most popular and most widely used model system for the study of the physical properties of biomembranes. Lipid membranes on a solid support were reported by McConnel et al. in 1988 [32]. Below we present the methods of formation and basic physical and structural properties of these systems. 3.3.1 Lipid Monolayers

Figure 3.9 A schematic representation of the cytoskeleton of erythrocyte membranes.

Lipid monolayers are formed spontaneously at the air–water interface. This is due to the amphiphilic nature

96

Bioelectrochemistry

Figure 3.11 (a) Expanded and condensed monolayer on a water surface. (b) Scheme for the deposition of monolayers onto a hydrophobic surface (e.g. mica). The immersed surface becomes hydrophilic after deposition of the first layer and becomes hydrophobic after deposition of the second layer.

of lipids. When lipids are dissolved in a non-aqueous volatile solvent and introduced onto a polar liquid surface, the solvent will evaporates leaving the lipid molecules oriented at the liquid–gas interface. The polar head groups pull the molecule into the bulk of the water and the hydrophobic chains are oriented into the air. Sweeping a barrier over the water surface causes the molecules to come closer together and to form a compressed and ordered monolayer (Figure 3.11a). The formation of thin oil films at an air–water interface was first reported in the 18th century by Benjamin Franklin. During his visit London in 1773, he observed that one teaspoon of oil spread on the water surface had a calming influence over half an acre (2000 m2) of water. Taking into account the volume of oil used (5 ml) this would mean that a film thickness 0.25 mm (about 100 layers) was covering the surface. Franklin reported his finding to the Royal Society of London in 1774. The investigation of the properties of oil films at an air–water interface was started by Agnes Pockels with a very simple trough in her kitchen. She reported her results in a letter to Lord Rayleigh. This letter was published in Nature in 1891. She was the first to perform experiments with monolayers using a barrier. Considerable progress in the physical study of monolayers was achieved thanks to the work of Irving Langmuir. Langmuir studied the relationship between the pressure and area on an aqueous surface. Further, Katherine Blodgett, who worked with Langmuir, developed the technique of transferring the films onto solid substrates (Figure 3.11b). A brief history of Langmuir–Blodgett films has been published by Gaines [33].

Below we consider the basic principles of monolayer thermodynamics and the properties of lipid monolayers at an air–water interface. The basic physical value that characterizes the lipid monolayer is the surface tension g. Surface Tension The changes in internal energy at a solid–liquid interface is characterized by Equation 3.1 [34] X dU ¼ TdS þ mi dni  PdV þ gdA ð3:1Þ

where U is the internal energy of the system, S is the entropy, mi and ni are the chemical potential and the number of moles of component i, respectively, A is total interfacial area and g is the surface tension of the interface. Since the Gibbs free energy G ¼ UTS þ pV, it follows that at constant p and using the surface excess quantities dGex ¼  Sex dT þ and g ¼ dGex =dAjT;p;ni

X

mi dnex i þ gdA

ð3:2Þ

ð3:3Þ

In the case of a pure liquid in equilibrium with its saturated vapor, the surface tension is also equal to the surface excess of the Helmohltz free energy (F ¼ G  pV) per unit area g 0 ¼ F ex 0 =A

ð3:4Þ

Biological Membranes and Membrane Mimics

97

mica, platinum or filter paper, is partially immersed in the liquid phase and is connected to an electromicrobalance. The forces acting on the plate are its weight Fp and surface tension effects downward, and Archimedes buoyancy Fa upward (Figure 3.12b). The net downward force is F ¼ F p þ 2gðw þ tÞcos Q  F a

ð3:6Þ

Where w and t (t  w) are the width and the thickness of the plane, respectively, and Y is the contact angle of the liquid with the solid plate. If the plate is completely wetted, the contact angle Y ¼ 0 and cos Y ¼ 1, so that F ¼ F p þ 2gw  F a

ð3:7Þ

When the composition of the interface varies, Fp and Fa (provided the plate is maintained in a fixed position) stay constant, and DF ¼ 2w (gsolutiongwater) ¼ 2wp and p ¼  DF=2w

ð3:8Þ

Properties of Lipid Monolayers Figure 3.12 (a) Langmuir–Blodgett trough with lipid monolayer. (b) Schematic representation of the principle of the measurement the surface pressure. For a description see the text.

To illustrate the physical meaning of the surface tension, let us consider a lipid film at an air–water interface of a Langmuir–Blodgett trough (Figure 3.12a). The lipid molecules in the insoluble monolayer can only move parallel of the water surface. The molecules hit against a movable barrier creating a surface pressure p. The work done in extending the movable barrier a distance Dx is: p l Dx. On the other side, the change in surface energy with the exchange of the lipid monolayer on a pure water surface is (g0  g)l Dx, where g0 and g are the surface tensions of the water and monolayer, respectively. Thus p ¼ Force=l ¼ g 0  g

ð3:5Þ

The surface tension is then a force per length unit expressed as N m1. As an example, the surface tension of pure water, g0 ¼ 72.75 mN m1 [35]. A fluid interface, such as an air–water interface, has the advantage of being a plane interface, the change in interfacial free energy of which can be obtained by simple measurement of surface pressure. The most common method for measurement of the surface pressure is the Wilhelmy method [36]. According to this method a thin plate, usually made of glass,

Phospholipids form stable monolayers at an air–water interface. The forces between the polar head groups are ionic and are proportional to 1/r (r is the intermolecular separation). The forces between hydrocarbon chains are due to van der Waal’s interactions and are proportional to 1/r6 (attraction forces) and 1/r12 (repulsive forces). Thus, the interactions in the sub-phase are of longer range than those in the super-phase. When the lipids are spread over sufficiently large surface area and no external pressure is applied to the monolayer, the molecules behave as a two-dimensional gas (Region G in Figure 3.13), which can be described by pA ¼ kT

ð3:9Þ

where p is the surface pressure, A the molecular area, k is the Boltzman constant and T is the thermodynamic temperature. With further compression, ordering of the film takes place and it behaves as a two-dimensional liquid. This state, the so-called liquid expanded state (L-E), is shown in Figure 3.13. With continued compression, the L-E state turns into a liquid condened state (L-C). Further compression results in the solid state (S). This solid state is characterized by a steep and usually linear relationship between the surface pressure and the molecular area. The collapse pressure pC is reached with further compression, at which the film irretrievably loses its monomolecular form. The forces exerted upon it become too strong for confinement in two dimensions and molecules are ejected

98

Bioelectrochemistry

Figure 3.13 (a) Compression isotherm of the monolayer: G – gas state; L-E liquid expanded state; L-C liquid condensed state; S solid state; C – collapse; pC collapse pressure; pt transition pressure (at the beginning of the L-E-LC transition); At mean area at pt, A0 – limiting area, AS, area in the solid state. The arrow indicates the direction of the compression. (b) Schematic representation of the monolayer structures in different states.

out of the monolayer plane into either the sub-phase (more hydrophilic molecules) or the super-phase (more hydrophobic molecules). However, collapse is not uniform across the monolayer, but is usually initiated near the leading edge of the barrier or at discontinuities in the trough, such as corners or the Wilhelmy plate. Usually a collapsed film will consist of large areas of uncollapsed monolayer containing islands of collapsed regions. The value of the collapse pressure varies depending on the phopsholipid structure and temperature. For simple saturated fatty acids the collapse pressure can be in excess of 50 mN m1 which is equivalent to about 200 atmospheres if extrapolated to three dimensions. The p-A isotherms for real monolayers can be well described by a two-dimensional analog of the van der Waal’s equation 



 a ðA  bÞ ¼ kT A2

cross the x-axis. This point corresponds to the hypothetical area occupied by one molecule in either the solid or liquid-condensed state (Figure 3.13). The shapes of the p–A isotherms of lipid monolayers depend on temperature. This is shown in Figure 3.14 where the plot of surface pressure as a function of area per molecule for a lipid monolayer composed of dimyristoylphosphatidylcholine (DMPC) is presented for two temperatures of water subphase. The isotherm labeled T2 represents monolayers in the liquid-crystalline state. In this case the monolayer does not exhibit phase transition. At lower tempretures, however, the phase transition from fluid to rigid film takes place (see also [37]). A phase transition in a monolayer can also take place at a constant pressure. For example, in a monolayer composed of myristine acid, the phase transition from a

ð3:10Þ

where a is the van der Waal’s constant, which characterizes the intermolecular interaction, and b is the effective area of the molecule cross-section (b  A0). Quantitative information can be obtained on the molecular dimensions and shape of the molecules under study. When the monolayer is in a two-dimensional solid (S) or liquid condensed state (L-C), the molecules are relatively well oriented and closely packed and the zero pressure molecular area can be obtaining by extrapolating the slope of the solid (AS) or liquid-condensed (A0) phase to zero pressure – the point at which these lines

Figure 3.14 The plot of surface pressure as a function of area per molecule for a monolayer composed of dimyristoyl-phosphatidylcholine (DMPC) at two temperatures of water subphase: T1 ¼ 20 and T2 ¼ 28  C.

Biological Membranes and Membrane Mimics

condensed to a liquid expanded state takes place over a narrow temperature interval (2  C). This transition is accompanied by an increase in the area per molecule, from 0.21 to 0.4 nm2 [18]. The calculation of the thermodynamic parameters of monolayers at phase transitions is discussed in [38]. The pA isotherms can be effectively used for estimation of the area per molecule of phospholipids and cholesterol. This area depends on both the structure and the charge of the head group, as well as on the length and the degree of saturation of the hydrocarbon chains. For example, a study of lipid monolayers composed of different phosphatidylcholines: dioleoyl phosphatidylcholine (DOPC) has both chains unsaturated, soy bean phosphatidylcholine (SBPC) has polyunsaturated fatty acids chains, egg phosphatidylcholine (eggPC) has 50 % saturated and 50 % unsaturated chains and dipalmitoylphosphatidylcholine (DPPC) has both chains saturated, revealed that at any fixed surface pressure the areas per molecule are in the following order: DOPC > SBPC > eggPC > DPPC. The corresponding intermolecular distance was calculated to be 1.06, 1.0, 0.97 and 0.81 nm at a surface pressure of 20 mN m1 [39]. Thus, a change in the saturation of the fatty acid results in changes in the intermolecular distance of the monolayer. The area per molecule is a sensitive indicator of the structure of amphiphilic molecules. This is illustrated in Figure 3.15, in which the p  A isotherms of lipid monolayers composed of cholesterol (CH), sphingomyeline (SM) and DOPC are presented. It can be seen that, at any fixed surface pressure, the areas per molecule are in the following order: DOPC > SM > CH. Figure 3.16 schematically illustrates the area per molecule and intermolecular distance for these compounds. If the monolayer is composed of a mixture of different phospholipids, then, depending on the structure of phospholipids, the monolayer could be less or more densely packed. Obviously, at a constant surface pressure and in the case of ideal miscibility or in the case of lack of miscibility, the plot of the average area per molecule as a function of concentration of one of the components should be a straight line A12 ¼ xA1 þ ð1  xÞA2

ð3:11Þ

where x is the mole fraction of component and A1, and A2 are extrapolated ‘zero-pressure’ areas for the corresponding phospholipids. Any deviation from linearity indicates changes in the miscibility of the monolayer components and can indicate formation of supermolecular assemblies or domains. These domains can be observable by fluorescence microscopy [40]. The lipid monolayers can also be

99

Figure 3.15 Surface pressure–area isotherms of monolayers formed by cholesterol (CH), Sphingomyelin (SP) and dioleoylphosphatidylcholine (DOPC) on a water sub-phase containing 10 mM NaCl þ 10 mM Tris-HCl, pH 7.4, T ¼ 25  C.

effectively used for the study of the mechanisms of protein–lipid interactions [36], and the functioning of phospholipases [41]. These are wide application of monolayers in connection with nanotechnologies. The Langmuir–Blodgett method of deposition of lipid monolayers on a solid support allows the preparation of biosensors composed of thin films, as well as the use of other powerful techniques for the study of physical and structural properties of thin films, such as Fourier transform infrared spectroscopy (FTIR), atomic force microscopy (AFM), scanning tunneling microscopy (STM) and other methods (see [42] for applications of lipid monolayers). The exact correspondence of the properties of lipid monolayers to the properties of biomembranes is, however, still under discussion [6] and is particularly connected with the selection of the monolayer surface pressure that most closely corresponds to that of bilayers, as well as with questions concerning the mechanisms of interaction between the two monolayers that create bilayers. Marsh suggested that p should be in the range 30–35 mN m1 [43]. However, the main transition temperature, TM, for DPPC then occurs about 5  C lower than for bilayers. Other authors suggest p ¼ 50 mN m1 [44]. This gives the correct TM, but the area per molecule at T ¼ 50  C for DPPC monolayers is less then that for bilayers [6]. The prediction of bilayer properties using monolayers would be most correct when no specific interaction exists between the two monolayers. There is, however, evidence that such interactions should exist [6]. Despite of this, there is an obvious advantage in the application of monolayers to study the

100

Bioelectrochemistry

Figure 3.16 Schematic representation of the areas per molecule and intermolecular distances for cholesterol (CH), Sphingomyelin (SP) and dioleoylphosphatidylcholine (DOPC) based on the data plotted in Figure 3.15.

surface properties of lipid membranes. Monolayers allows much easier measurement of the area per phospholipid then bilayers. It is also easy to measure the dipole potential of monolayers and to study the adsorption processes at the water–monolayer interface [44]. These advantages, together with other new applications in nanotechnology, shows that lipid monolayers represent a very attractive area of study for physics as well as for bioelectrochemistry. The study of the thermodynamic properties of lipid monolayers can be performed by using precise Langmuir–Blodgett troughs (e.g. NIMA Technology Ltd. produces troughs of different sizes including tensiometers and dippers for the deposition of films on a solid support [46]).

Dp ¼ gð1=R1 þ 1=R2 Þ

ð3:12Þ

where R1 and R2 are the inner and outer radii of the surface curvature and g is the surface tension. In the central part of the membrane, the radius of curvature is close to infinity, i.e. R1 ¼ R2 ! 1. Therefore the pressure difference is close to zero: Dp ¼ 0. However, the pressure at the central part of the Plateau–Gibbs border is lower then that in the water phase, i.e. Dp < 0. Therefore the solvent will move from the thin, flat part of the BLM to the Plateau–Gibbs border. This will cause further thinning of the membrane (Figure 3.17). This process can also be observed visually in reflected light. Thick films are colored, like oil films on a water surface. As soon as the films become thinner, black

3.3.2 Bilayer Lipid Membranes (BLM) Formation and Electrical Properties of BLMs Stable bilayer lipid membranes (BLM) were first reported in 1962 by Mueller and coworkers [30]. Due to the amphiphilic nature of phospholipids, they spontaneously form lipid bilayers in a water phase. In experiments by Mueller et al., a BLM was formed from a crude fraction of phospholipids in circular holes of relatively small diameter (0.8–2.5 mm) in a Teflon cup immersed in a larger glass compartment filed by electrolyte (Figure 3.17). Small amounts of phospholipid dissolved in hydrocarbon solvent (e.g. n-heptane or n-decane) can be placed into the hole using a simple brush or a Pasteur pipette. Immersion of the drop of phospholipid in the hole results in immediate formation of a relatively thick lipid film (thickness > 100 nm). The behavior of this thick film is determined by differences in the hydrostatic pressures in the peripheral part (Plateau–Gibbs border) and in the central part, which is relatively flat. According to Laplace law, the difference in the hydrodynamic pressures between the inner and outer phases is determined by equation (3.12)

Figure 3.17 (a) The scheme of current measurement across a BLM using a current to voltage converter and a diagram of the formation of the BLMs. (b) The electrical scheme for the system BLM–electrolyte–electrodes, Rm – membrane resistance, Cm – membrane capacitance, Re – resistance of electrodes and electrolyte.

Biological Membranes and Membrane Mimics

spots start to appear. These membranes were therefore initially called ‘black lipid membranes’ with the abbreviation BLM. Currently this abbreviation is used also for ‘bilayer lipid membrane.’ Observation of thin film formation, shows that the black spots form non-uniformly and unsymmetrically. As soon as the films become thinner van der Waal’s forces also start to play a considerable role in film formation. The resulting BLM is characterized by a bilayer part surrounded by a Plateau–Gibbs border (Figure 3.17). The formation of a BLM can be easily detected by measurement of the electrical properties of the film – conductance and capacitance. BLM conductivity can be easily measured by a current to voltage converter (Figure 3.17). The dc voltage (of a relatively small amplitude U0 ¼ 50–100 mV) is applied to the BLM through e.g. Ag/AgCl electrodes (usually immersed into salt agar bridges). One electrode is connected to the high resistance input (usually 1012 O) of the operational amplifier (e.g. Analog Devices AD 548J). The current i flowing through the feedback resistor Ri is transformed into the output voltage U according to the equation: i ¼ U/Ri. Thus, measuring the output voltage by, for example, a millivoltmeter, chart recorder or connecting the output of the amplifier in to the A/D converter and then to a computer, it is possible to determine the current i and its changes with time. Knowing the amplitude of the applied voltage U0, it is then possible to determine the membrane’s specific conductance g ¼ i/ U0A ¼ U/U0RiA (A is the area of the BLM, which can be determined using a microscope). (The principles of operation of amplifiers with negative feedback are described in [47].) Experiments to study membrane conductivity can be performed cheaply using a simple apparatus based on an operational amplifier. Care should be taken to properly shield the measuring chamber, due to the rather small current that is flowing through the BLM (typically of the order of pA). It is, however, possible to use commercial instruments (e.g. Keithley 6512 (USA)) which can be connected on line with a PC through the KPC-488.2AT Hi Speed IEEE-Interface board. A BLM is characterized by rather a low specific conductivity, which is in the range 106–1010 S cm2, with a typical value being about 108 S 1 cm2 [18]. This phenomenon is connected with the low dielectric permittivity of the inner, hydrophobic part of the membrane. The relative dielectric permittivity of this part is similar to that characteristic of n-alkanes: 2.1. The mechanisms of conductivity of BLMs will be discussed below (see Section 3.9). BLM electrical capacitance is an important value that characterizes the dielectric and geometric properties of the membrane. Usually it is the specific capacitance that is

101

used for characterization of the BLM C S ¼ C=A ¼ ee0 =d

ð3:13Þ

where A is the area of the bilayer part of the BLM, d is the thickness of the hydrophobic part of the BLM, e  2.1 is the relative dielectric permittivity and e0 ¼ 8.85 · 1012 F m1 is the permittivity of free space (vacuum). The simple electrical model of the membrane can be represented by a capacitor connected in parallel with a resistance. The resistance of the electrolyte and measuring electrodes can be taken into account as a resistance connected in series (Figure 3.17b). The membrane capacitance can be measured by an alternating current bridge [48]. However, the most precise measurement of conductance and capacitance can be performed by measurement of the complex impedance of the BLM [48,49]. The theory of the impedance is described in detail in many textbooks, monographs and reviews, see for example [48,49]. Here we restrict ourselves to consideration of the basic phenomena. Impedance measurements are made by applying a small alternating (a.c.) current of known circular frequency o and a small amplitude i0 to a system, and measuring the amplitude U0 and the phase difference j of the electrical potential. The impedance is usually represented by the absolute values of the impedance and the phase j Z~ j ¼ U 0 =i0

and ff Z~ ¼ j

ð3:14Þ

In cartesian coordinates, impedance becomes a complex number Z~ ¼ R þ jX;

where j ¼

pffiffiffiffiffiffiffiffi 1

ð3:15Þ

~ describe the resistance The real and imaginary parts of Z (R) and reactance (X), respectively, and can be represented by appropriate electrical circuit elements. In the case of an unmodified BLM, i.e. when the membrane resistance is considerably higher than the absolute value of reactance, the equivalent electrical circuit can be simplified and reduced to a resistance in series with a capacitance (Figure 3.18c). For this circuit the complex impedance is Z~ ¼ R  j=wC

ð3:16Þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 ~ j ¼ R þ1=ðoCÞ2 The absolute value of impedance is j Z ~ j and j and the phase j ¼ arctg(1/(oRC)). The values of j Z are usually presented as a function of frequency, i.e. the so~ j should decrease with called Bode plot. In this plot j Z increasing frequency and approach R at high frequencies (Figure 3.18a). At lower frequencies, the phase j is equal to

102

Bioelectrochemistry

Figure 3.18 Impedance spectra of the BLM formed from diphytanoyl phosphatidylcholine in n-decane in 10 mmol l1 KCl. Bode plot of the absolute value of: (a) impedance; (b) phase. The points are experimental values and line represent fit according to the equivalent electrical circuit (c).

90 as for a pure capacitor, but at higher frequencies it decreases dramatically, approaching zero as for a pure resistance (Figure 3.18b). The typical Bode plot for an unmodified BLM made from diphytanoyl phosphatidylcholine (Avanti Polar Lipids, Inc, USA) (lipid was dissolved in n-decane (Fluka) to a concentration of 30 mg ml1) is presented in Figure 3.18. The points are the experimental values measured by an electrochemical analyzer, Autolab PGSTAT 12, equipped with FRA 2 module (Eco Chemie, The Netherlands) and the line represents the fit according to the equivalent circuit (Figure 3.18c). The BLM was formed in a circular hole of diameter approx. 0.8 mm in 10 mmol l1 KCl. The typical values for R (i.e. the resistance of electrolyte and electrodes) and C (i.e. the membrane capacitance) were 15 kO and 1.67 nF, respectively. If the BLM is modified by compounds, e.g. channel formers or carriers, its conductivity increases and the equivalent electrical circuit is more complex [49,50]. The application of impedance spectroscopy allows us not only to obtain the basic electrical parameters of the BLM, but also to analyze the mechanisms of ionic transport (see for example [50] and references therein). Currently, there are several pruducers of sophisticated electrochemical or impedance analyzers that can be successfully used for the study the impedance of BLMs. Among others, the already mentioned Autolab PGSTAT 12 with FRA 2 module is a universal high quality instrument allowing, not only the measurement of impedance spectra with high sensitivity over a wide range of frequencies (1 mHz to 1 MHz), but also the complete electrochemical parameters of the system. Solartron (UK), BAS-Zahner (USA) and CH Instruments Inc. (USA) produce electrochemical and impedance

analyzers that can also be used for the study of BLM electrical properties. The specific capacitance and the thickness of a BLM depend on the content of organic solvent (e.g. n-heptane or n-hexane). In addition, using the solvent with larger hydrocarbon chains, e.g. n-hexadecane or squalene [51,52], it is possible to obtain thinner membranes. This is due to the fact that sterical repulsion squeezes the larger molecules of solvent out of the bilayer to its border part. There have been various methods developed to make thinner membranes and to study the properties to the bilayer of a biomembrane, which obviously does not contain solvent (except minor content of fatty acids that appear due to the action of phospholipases). In [51] these methods are described in detail. In addition to the approach using hydrocarbon solvents with a larger length of hydrocarbon chains, it is also possible to freeze out the solvent, i.e. n-hexadecane. This solvent is characterized by a phase transition temperature from the liquid to the solid state below approx. 15.5  C. Therefore, if the surrounding electrolyte is slowly cooled, it is possible to crystallize the solvent, which is moved from the bilayer to the Plateau–Gibbs border of the membrane [53]. The method of ‘drying’ the membrane developed by Rovin [54], is based on using two types of solvent for formation of the BLM dioxane and n-octane, in different ratios. Dioxane is a solvent which is miscible with both water and n-alkanes. Therefore, if the BLM is formed from a mixture of lipids and a mixture of solvents, one of which is dioxane, the dioxane will be moved out of the bilayer to the surrounding electrolyte, and the membrane specific capacitance will increase.

Biological Membranes and Membrane Mimics

In 1972 Montal and Mueller [55] proposed a method of formation BLMs from monolayers. According to this method, a Teflon cup divided by a wall with a circular hole of diameter approx. 0.3 mm is filled with electrolyte to just below the bottom of the hole. Then a small amount of lipid dissolved in chloroform is added to the water surface in both compartments of the Teflon cup. After the chloroform is evaporated (approx. 10 min) the level of the water phase is increased by slow addition of electrolyte using a syringe. As soon as the level of the electrolyte surpasses the top of the hole, the BLM is formed. It should be noted that the membrane is not entirely without solvent, because for membrane stability it is necessary that the Plateau–Gibbs border is present. This is attained by the addition of a small amount of lipid dissolved in n-hexadecane to the hole. Other method of formation of BLMs containing proteins was proposed by Schindler [56]. This method is a modification of that proposed by Montal and Mueller [55] and consists in a different formation of the monolayers. For this purpose liposomes or proteoliposomes are added to both compartments of the Teflon cup. The lipid monolayers are formed due to the fact that liposomes, when in contact with a hydrophobic–air interface with a low surface pressure, are broken and the lipids form a monolayer at the air–water interface. The process of BLM formation is analogous to that of Montal and Mueller. Using the various methods of BLM formation it is possible to considerably change the thickness and the specific capacitance of BLMs as is shown in Table 3.4. With increasing hydrocarbon chain length of the solvent, the membrane thickness decreases independently of the lipid composition or cholesterol content. The lipid composition only affects the changes in thickness with changes in the number of carbons of n-alkanes [18]. Thus, by varying the hydrocarbon solvent and the type of phospholipids, it is possible to obtain to BLMs with the desired thickness.

103

Table 3.4 The specific capacitance and the thickness of a BLM formed from dioleoylphosphatidyl choline (DOPC) using the method of Mueller et al. [30] and n-alkanes of various length, as well as that for solvent-free BLMs prepared according to the method of Montal and Mueller [55]. Number of carbon atoms in n-alkane

Specific capacitance, 103 F m2

Thickness, nm

Reference

8 10 12 14 16 Solvent free

3.77 3.74 4.22 4.86 6.24 7.28

4.93 4.97 4.46 3.82 2.98 2.56

[57] [57] [57] [57] [58] [57,58]

The electrical and other physical properties of BLMs formed from natural phospholipids are similar to that of biomembranes (Table 3.5). Unmodified BLMs are characterized by low conductivity. They do not reveal any metabolic activity and are not selective for transport of ions, unlike biomembranes. However, BLMs can be modified by channel formers, carriers or receptors, which provide sensitivity and selectivity of ionic transport or ligand–receptor interactions, as with biomembranes. Also in the presence of various modifiers, the conductivities of BLMs usually increase. These effects, together with several similar parameters show that the properties of lipid bilayers are close to those of biomembranes. Therefore the study of the physical properties of BLMs has significance for understanding the properties of biomembranes. Stability of BLMs: Electrical Breakdown and Electroporation The main problem in the application of BLMs to electrochemical studies is their relatively low stability. Even in

Table 3.5 Comparison of the properties of BLM and biomembranes Properties

Biomembranes

BLM

Thickness, nm Surface tension, mN m1 Conductivity, O1 cm2 Specific electrical capacity, 103 F m2 Breakdown voltage, mV Refractive index Permeability for water, mm s1 Energy of activation for water permeability, kJ mol1 Ionic selectivity, PKþ/PNaþ

6.0–10.0 0.03–3.0 102–105 5–13 100 1.6 0.5–400 40.3 1–25

2.5–8.0 0.2–6.0 106–1010 2–10 150–300 1.56–1.66 31.7 53.3 5.4–9.0

104

Bioelectrochemistry

the case of a relatively low potential differences across the BLM comparable with those found in biomembranes, i.e. DU ¼ 0.1 V, due to small thicknesses (d  108 m), rather large electrical fields exist across the membrane: E ¼ DU/h  107 V m1. Thus, even a small decrease in the membrane thickness, e.g. due to thermal fluctuations, results in an increase in the electrical field that can induce the electrical breakdown of the membrane. The theory of the electrical breakdown of BLMs was developed mostly due to work by Chizmadzhev and coworkers [59]. They found that the mean lifetime of a BLM in an electric field decreases with increasing voltage difference across the BLM. The membrane breakdown is connected with the appearance of conductive pores in the lipid bilayer due to membrane electrostriction, i.e. the compressive force F ¼ eme0U2/ (2d2), which originates from the applied voltage, U [60]. Let us consider the changes in the energy of the membrane if a cylindrical amphiphilic pore with a radius r appears. The changes in the energy of the pore can be described by the Equation (3.17) E ¼ 2prg l d  pr 2 g  DCU 2 =2

ð3:17Þ

where d is the membrane thickness, r is the radius of the pore, gl is the coefficient of linear tension of the pore, g is the surface tension of the membrane, DC is the change in electrical capacitance and U is the potential difference across the membrane. The first term in Equation (3.17) is connected with the changes in the energy of the membrane due to the appearance of the cylinder surface between the membrane and the pore. The second term denotes the decrease in surface energy due to the decrease in the membrane surface, which is equal to the cross-sectional area of the cylinder. The third term is connected with changes in the energy of the capacitor due to changes in the dielectric permittivity of the membrane interior (the dielectric part with a dielectric permittivity em  2.1 is replaced by water with dielectric permittivity ew  80). The changes in electrical capacitance can be expressed as: DC ¼ pr2Cs(ew/em  1) where Cs is the specific capacitance of the membrane, i.e. the capacitance per unit area. Equation (3.17) can therefore be transformed to E ¼ 2prg l d  pr 2 ðg þ CU 2 =2Þ

ð3:18Þ

where C ¼ Cs(ew/em  1). As soon as the pore radius increases, its energy should change non-monotonously and can be described by a curve with a maximum as displayed in Figure 3.19a. It is clear from the figure that a membrane defect with a small radius has a tendency to diminish in size.

However, pores with a radius larger than a certain critical value rc will increase irreversibly and cause the membrane to break down. The value of the critical radius can be found by differentiating the energy dE/dr ¼ 0 and is described by Equation (3.19) [61] rc ¼ g l d=ðg þ CU 2 =2Þ

ð3:19Þ

The value of rc is of the order of the membrane thickness, i.e. several nm. After substitution of Equation (3.19) into Equation (3.18) we can obtain the dependence of the energy of the pore on the membrane potential E ¼ ðg l dÞ2 =ðg þ CU 2 =2Þ

ð3:20Þ

Thus, application of a potential to a membrane will result in a decrease in the potential barrier and the probability that the membrane will break down will increase (Figure 3.19a). Above we considered a simplified situation, where the breakdown of the membrane is connected with formation of amphiphilic pore. As a matter of fact, owing to thermal fluctuations of lipid molecules, hydrophobic pores are formed prior to hydrophilic ones (Figure 3.19c). When a hydrophobic pore exceeds this critical radius, a reorientation of the lipids converts the pore into a hydrophilic one (Figure 3.19c) [62]. The dependence of the energy of pore formation on the radius is, in this case, more complicated (Figure 3.19b). The study of BLM breakdown has great significance for understanding processes, such as membrane fusion, lysis and apoptosis of cells, that may involve the opening of a lipid pore to combine volumes initially separated by membranes. Experimental studies have revealed two types of BLM behavior under electrical stress: reversible and irreversible electrical breakdown. Irreversible breakdown is observed for membranes of any lipid composition. It is accompanied by a rapid increase in membrane conductance and results in mechanical breakdown of the BLM [63]. However, for BLMs of a specific lipid composition, e.g. oxidized cholesterol, or at the phase transition of phospholipids [64], reversible pores are observed when the membrane is exposed to a short pulse of high electric field. In this case, even after a five to six order increase in the conductance, it then drops to the initial level upon voltage decrease [62,65]. It has been assumed, that in the case of irreversible breakdown, few pores are formed before the first of them reaches a critical radius and starts irreversible expansion leading to membrane rupture. In the case of reversible breakdown, a large population of pores accumulates under high voltage

Biological Membranes and Membrane Mimics

105

Figure 3.19 The dependence of the energy of a pore in a membrane on its radius in the case of: (a) a hydrophilic pore; (b) a hydrophobic pore formed first and then transformed into a hydrophilic one at a certain critical radius; (c) a schematic representation of the transformation of a hydrophobic pore into a hydrophilic one [68]. (Reprinted from Bioelectrochem. Bioenerget., Weaver and Chizmadzhev, Theory of electroporation: a review, 41, 135–160, 1996, with permission from Elsevier.)

before the BLM ruptures. Recently it was observed that application of a voltage (150–500 mV) to a BLM results in fast transition between different conductance levels reflecting opening and closing of metastable pores [66]. The mean lifetime of the pores was 3 ms at U ¼ 250 mV, however pores with longer lifetimes, up to 1 s were observed as well. Based on the conductance value and its dependence on the ion size, the radius of the average pore of a 0.5 nS conductance was estimated as 1 nm. This pore might involve only 100 lipid molecules, which corresponds to less then 108 % of all the lipids in a BLMs of 1 mm2 area. The metastable lipidic pores identified in this study most probably correspond to the small metastable lipid pores whose existence was assumed earlier to explain the accumulation of a very large number of pores during reversible electroporation of BLMs modified by uranyl ions [62]. The physical structure of non-conducting pre-pores remains unknown. Opening and expansion of conductive hydrophilic pores, i.e. pores with the edges formed by polar head groups of phospholipids, is assumed to be preceded by formation of very small and short lifetime hydrophobic pores with the edges formed by the lipid hydrocarbon chains. Evolution of hydrophobic pore into hydrophilic pores involves reorientation of the polar heads of the lipids from the surface of the bilayer to the edge of the pore (Figure 3.19). Melikov et al. [66] assumed that the pre-pores correspond to small clusters of lipids with their polar heads trapped inside the hydrophobic interior of the membrane upon closing of a

hydrophilic or partially hydrophilic pore. Interaction between the polar lipid heads in the same cluster can increase the lifetime of the cluster and thus stabilize the pre-pore state. Alternatively the pre-pore state may correspond to a cluster of water molecules trapped inside a hydrophobic interior. Such clusters of lipid polar heads or water molecules will then transform back into a small hydrophilic pore. Changes in the conductance of a membrane following the application of external voltage, i.e. electropermeability, has great practical significance. By means of the application of an external field it is possible to incorporate DNA or drugs into cells. This effect, known as electroporation, has been known for several decades and has been reviewed in several papers (see, for example, [67,68]). This process has been studied in most detail using BLMs [68]. It has been shown that charged ions or small molecules can be driven through the BLM by electromigration. The mechanism of electroporation of larger molecules, like proteins, is, however, not yet clear. The peculiarities of the electropermeability of tissues, e.g. human skin, under the application of electrical field revealed similar behavior to that of the BLMs. However, the theory of electroporation of these complex structures remains still incomplete (see, for example [69]). For a deeper look into the problem of electroporation on a molecular level, we recommend the extensive review by Weaver and Chizmadzhev [68].

106

Bioelectrochemistry

3.3.3 Supported Bilayer Lipid Membranes As we already mentioned in the section devoted to the lipid monolayers, using the Langmuir–Blodgett technique, it is possible to obtain supported lipid membranes. A simple method of formation of a lipid film on a metal support was proposed by Tien and Salamon [70]. A silver wire of approx. 0.3 mm diameter coated in Teflon, is immersed into a lipid solution in n-decane. Then the tip of the wire is cut by a sharp knife and immediately immersed into the electrolyte, where the formation of the thin film occurrs spontaneously. The disadvantage of this method consists in the fact that the film, formed on a rather rough metal surface, is inhomogeneous and is composed of monolayers, bilayers or even multilayers [71]. Application of dc voltage during film formation results in improvements of the film characteristics and the membrane becomes more homogeneous [72]. However, the most homogeneous films can be obtained using a smooth gold surface with chemisorbed alkanethiols. According to this method, the surface of the gold is cleaned and then immersed in a solution of

alkanethiol, e.g. hexadecane thiol. Thiols provide almost covalent binding of the alkanethiol to gold and the hydrophobic chains of hexadecane thiols will create a densely packed hydrophobic surface (Figure 3.20a). The tilting of the chains away from the surface normal occurs because the spacing of the three-fold hollow sites on the Au (1 1 1) surface, into which the –SH head groups fit, is slightly larger than the optimal van der Waals distance between adjacent hydrocarbon chains: by tilting at an angle of 20–25  , the chains adjust their spacing to optimize the van der Waals interaction [73]. In an open circuit the formation of an alkane thiol film takes about 12 hours. However, application of a dc voltage with amplitude 600 mV (positive terminal on a gold) results in the fast formation (a few minutes) of a homogeneous alkanethiol monolayer [74]. A second monolayer can be formed, e.g. by the Langmuir–Blodgett technique, by immersion of the gold electrode into the lipid solution or by liposome fusion [75]. The supported lipid membranes (sBLMs) are considerably more stable than BLMs [76–78]. While the breakdown voltage of a BLM

Figure 3.20 Schematic representation of sBLM and tBLM: (a) sBLM formed by chemisorption of alkylthiols on a gold. tBLMs are characterized by the hydrophilic spacer between the gold and the lipid bilayer which can be formed by specially synthesized lipids with a hydrophilic spacer (b), or by a lipopeptide layer (c) and/or a polymer layer (d).

Biological Membranes and Membrane Mimics

is less then 300 mV, that for an sBLM can reach more then 1 V [75]. The disadvantage of sBLMs consists in the fact that the alkanethiol monolayer is closely adjacent to the gold. Therefore it is impossible to incorporate, for example, large integral proteins into these membranes and also impossible to use these membranes for the study of the mechanism of ionic transport due to the lack of a water phase between the bilayers and the gold. This drawback has been solved by development of so-called tethered membranes (tBLM). tBLMs are similar to sBLMs, however, instead of alkanethiols, specially synthesized molecules with a hydrophilic spacers are used for the formation of a monolayer tethered to a gold support (Figure 3.20b). The hydrophilic space between the gold and the BLM can also be provided by lipopeptide layer (Figure 3.20c) and/or by a polymer layer (Figure 3.20d) [78,79]. sBLMs and tBLMs, due to their high stability and the unique properties that mimic a real biomembranes, can be used in nanotechnology, especially in the development of sensitive sensor systems (see Section 3.11). 3.3.4 Liposomes Liposomes (vesicles) are widely used models of the lipid bilayer of biomembranes. The formation of liposomes was first reported by Bangham [31]. They are formed from water dispersions of lipids by various methods. Depending on the method of formation one can obtain multilamellar or unilamellar liposomes of different sizes. The application of liposomes as a model of biomembranes has been discussed in a large number of reviews and monographs, see for example [80]. Methods of liposome preparation are considered in [81]. Multilamellar liposomes are the simplest particles in respect of preparation. Typically the desired amount of the lipid is dissolved in chloroform or in a mixture of chloroform and methanol. This mixture is allowed to evaporate under a stream of nitrogen in a spherical glass vessel. A rotary evaporator is usually used for this purpose in order to make a thin film of the lipids on the large surface of the glass wall. After the thin film is formed, it is then hydrated by addition of the desired volume of water or buffer. The lipid is then dispersed in water by vigorously vortexing for several minutes at a temperature higher than the phase transition temperature of the phospholipids. The concentration of lipids varies from method to method, but typically is a few mg ml1. For example, precise DSC calorimetry requires a lipids concentration of about 0.5 mg ml1, but densitometry requires larger concentrations – around 5 mg ml1. Multilamellar liposomes are rather large and their diameter is several

107

mm. The advantage of these liposomes is that they consist of a large numbers of bilayers and are characterized by a high degree of co-operativity in comparison with unilamellar liposomes. This co-operativity is expressed, for example, by narrower range of phase transition temperature. The disadvantage of multilamellar liposomes consists in their size inhomogeneity as well as in relatively fast sedimentation. They are therefore not suitable in experimental arrangements that do not allow stirring of the solution. Unilamellar liposomes can be prepared by various methods. The simplest one consists in sonication of the multilamellar liposomes with ultrasound [82] in an ultrasonic bath. This method results in formation of relatively small liposomes of approx. 20 nm diameter. The disadvantage of this method is the non-uniform diameter of the liposomes, as well as the possible presence of a certain amount of multilamellar liposomes. In addition, application of ultrasound could result in damage to the lipids. As a result the free fatty acids could also appear in a solution. Other methods are based on fast injection of an ethanol solution of lipids into a buffer [83] or by dialysis of water dispersions of micelles composed of lipids and detergents [84]. A rather useful method was proposed by McDonald et al. [85] that consists in formation of vesicles by extrusion of multilamellar vesicles through polycarbonate films. Depending on the size of the pores in a film, liposomes of the desired diameter can be prepared, usually from 50 nm to several mm. Commercial kits for the preparation of unilamellar liposomes are currently available, e.g. Avestin Inc. (Canada). Liposomes prepared by extrusion methods are rather homogeneous in size. Liposomes can be modified with various compounds, e.g. peripheral or integral proteins. In the case of integral proteins, the liposomes are prepared from a water dispersion of lipids containing the desired concentration of proteins (see, for example, [86,87]). 3.4 ORDERING, CONFORMATION AND MOLECULAR DYNAMICS OF LIPID BILAYERS As we have already mentioned, the lipid bilayers are liquid crystals of a smectic type. They are characterized by relatively fast lateral diffusion of lipids and about 600 times slower diffusion of proteins (the typical diffusion coefficients for lipids and proteins are: DLIP ¼ 6 · 1012 m2 s1, DPROT ¼ 1014 m2 s1, respectively). In the direction perpendicular to the membrane plane, the lipid bilayer is characterized by a certain ordering that depends on the conformation of the hydrocarbon chains of the phospholipids. This conformation depends on temperature and the phase transitions the lipid bilayer undergoes.

108

Bioelectrochemistry

3.4.1 Structural Parameters of Lipid Bilayers Measured by X-ray Diffraction The basic structural parameters of membranes can be determined by X-ray or neutron diffraction methods. X-rays are electromagnetic radiation with typical photon energies in the range 100 –100 keV. For diffraction applications, only short wavelength X-rays (hard X-rays) in the ˚ (intensities vary range of a few angstroms to 0.1 A between 1 and 120 keV) are used. Because the wavelengths of X-rays are comparable to the size of atoms and molecules, they are ideally suited for probing the structural arrangements of atoms and molecules in a wide range of materials. The energetic X-rays can penetrate deep into materials and provide information about the bulk structure [88]. X-rays are generally produced by either X-ray tubes or synchrotron radiation. In recent years synchrotron facilities have become widely used as the preferred sources for X-ray diffraction measurements. Synchrotron sources are thousands to millions of times more intense than laboratory X-ray tubes. X-rays primarily interact with electrons and atoms. Diffracted waves from different atoms can interfere with each other and the resultant intensity distribution is strongly modulated by this interaction. If the atoms are arranged in a periodic fashion, as in crystals, the diffracted waves will consist of sharp interference maxima (peaks) with the same symmetry as in the distribution of atoms. Measuring the diffraction pattern therefore allows one to deduce the distribution of atoms in a material. The peaks in a X-ray diffraction pattern are directly related to the atomic distances. Let us consider an incident X-ray beam interacting with the atoms arranged in a periodic manner, as shown in Figure 3.21. The atoms are located periodically in parallel planes. For a given set of lattice planes with an interplane distance of d, the condition for a diffraction (peak) to occur can be simply written as 2d sin Q ¼ nl

ð3:21Þ

which is known as Bragg’s law after W. L. Bragg, who first proposed it. In Equation (3.21), l is the wavelength of the X-ray, Y the scattering angle, and n is an integer representing the order of the diffraction peaks. Bragg’s law is one of most important laws used for interpreting X-ray and neutron diffraction data. In contrast to X-rays, neutrons have one great advantage, because deuteration dramatically changes the scattering of neutrons. Specific deuteration of component parts of the lipid, such as a selected methylene, therefore provides a localized contrast agent that leaves the system physically and chemically nearly equivalent [89]. The

Figure 3.21 Schematic representation of diffraction X-ray beams on two planes of a two-dimensional atomic lattice.

disadvantage of neutrons to X-ray diffraction is that neutron beams are much weaker and there are fewer sources of neutrons. In analogy with classical X-ray analysis, it is assumed that determination of bilayer structure means applying crystallography. It is however important to note that fully hydrated bilayers, even in a gel condition, are in far from a crystalline state. The contrast is more remarkable for a fluid La phase, where the hydrocarbon chains are conformationally disordered. The differences between crystalline structures and hydrated bilayers are particularly due to the high content of water, which allows for increased fluctuations. Therefore, due to fluctuations, it is impossible to determine the structures of biomembranes at an atomic level. However, bilayers in multilamellar vesicles (MLV), which are most often used in diffraction studies, are isotropically oriented in space and therefore give socalled powder patterns. The term ‘powder’ really means that the crystalline domains are randomly oriented in the sample. Therefore when the two-dimensional diffraction pattern is recorded, it shows concentric rings of scattering peaks, corresponding to the various d spacings in the crystal lattice. MLVs are characterized in a variety of sizes. Each bilayer is influenced by its neighbors. It is assumed that MLVs are ‘onion like,’ consisting of closed concentric spheres. Because lipid exchange between bilayers and solvent is slow, it is likely that the number of lipids in each bilayer remains constant over a fairly long time [6]. The advantage of MLVs is that they do not have to be especially oriented in an X-ray beam (as a matter of fact, they cannot be oriented). However, only a small fraction of the lipid in a powder sample diffracts from a given beam, so intensities are weak. In addition, due to short-range fluctuations (intrinsic fluctuations related to a single bilayer) and especially due to long-range fluctuations (fluctuations in the relative positions of unit the cell [90] (Figure 3.22)), the electron

Biological Membranes and Membrane Mimics

109

Figure 3.22 Schematic representation of long-range fluctuations in an MLV [6]. (Reprinted from Biochim. Biophys. Acta, 1469, Nagle and Tristram-Nagle, Structure of lipid bi-layers, 37. Copyright 2000, with permission from Elsevier.)

density profiles obtained by X-ray diffraction are broad. The typical electron density profile obtained for a MLV composed of DPPC is shown in Figure 3.23 together with a schematic representation of bilayer regions corresponding to head groups and hydrocarbon chains. The full thickness of the layer is composed of the thickness of the bilayer region DB and the thickness of the water region DW: D ¼ DB þ DW. DB is defined as: DB ¼ 2VL/A, where VL is the volume of lipid and A is its area. DW ¼ 2nWVW/A, where VW is volume of the water molecule and nW is the number of water molecules/lipid. The volume VL of the lipid is divided into head group DH and hydrocarbon chain regions DC: DB ¼ 2(DH þ DC). DC includes the hydrocarbon chain carbons except for the carbonyl carbon, which has substantial hydrophilic character. For DPPC the hydrophobic core therefore consists of 14 methylenes and one terminal methyl on each of two chains. The determination of the DC value is important for analysis of the mechanism of protein–lipid interactions, where the interactions between hydrophobic parts of the proteins and the lipids play a considerable role. The head group of the phospholipid consists of the remainder of the lipid. An electron density profile provides a good measure for the location of the phosphate groups. Information about z-co-ordinates of other groups has been obtained using neutron diffraction. Because of the thermal fluctuations, the position of the atoms in the lipid molecule can be described by a broad statistical distribution function, which can be obtained by molecular dynamics simulations (MD) (Figure 3.23) [91]. MD becomes rather attractive because

they allow one to obtain much greater detail than can be obtained experimentally. This detail can even be a guide for interpretation of experimental results. The latest structural parameters for fully hydrated lipid bilayers of the most frequently used composition are shown in Table 3.6. 3.4.2 Interactions between Bilayers We mentioned above that long-range fluctuations of bilayers cause less precision in the determination of the structures of bilayers by X-ray diffraction method. These fluctuations cause interactions between bilayers. However, two bilayers close to each other fluctuate less then those at a distance. When bilayers are at a close distance, then suppression of fluctuation takes place and results in a decrease in the entropy. This causes an increase in the free energy of the system. According to Helfrich [92], the free energy of the fluctuations increases with decreasing separation distance DW0 and depends on the bending elasticity modulus Kc of the bilayers F f l ¼ 0:42

ðkTÞ2 0 2 DW Kc

ð3:22Þ

The mechanisms of interbilayer interactions can be studied by X-ray diffraction [6] or by the surface force method [93]. X-ray diffraction studies are based on the measurement of the changes in the bilayer structural parameters following application of osmotic pressure. It

110

Bioelectrochemistry

Figure 3.23 Representation of the structure of DPCC in the La fluid phase: (a) probability distribution functions for different component groups from MD simulations [91]; the downward pointing arrows show the peak locations determined by neutron diffraction with 25 % water [89]; (b) electron density profile from X-ray studies; (c) A schematic representation of the head group and the hydrophobic region of the bilayers. The version on the left is a simple three compartment representation. The version on the right is a more realistic representation of the interfacial head group region. Dc is the experimentally determined Gibbs dividing surface for the hydrocarbon region [6]. (Reprinted from Biochim. Biophys. Acta, 1469, Nagle and Tristram-Nagle, Structure of lipid biolayers, 37. Copyright 2000, with permission from Elsevier.) Table 3.6 Structural parameters of fully hydrated lipid bilayers. Lipid Temperature

DPPC 20  C

DPPC 50  C

DMPC 30  C

DOPC 30  C

EPC 30  C

DLPE 20  C

DLPE 35  C

VL (Å3) D (Å) A (Å2) VC (Å3/region) VCH3 (Å3/group) 2DC (Å) DHH (Å) DB (Å) DW (Å) DH0 (Å) DB0 (Å) DW0 (Å) nW nW0

1144 63.5 47.9 825 25.9 34.4 44.2 47.8 15.7 9.0 52.4 11.1 12.6 3.7

1232 67 64 913 28.7 28.5 38.3 38.5 28.5 9.0 46.5 20.5 30.1 8.6

1101 62.7 59.6 782 28.1 26.2 36.0 36.9 25.8 9.0 44.2 18.5 25.6 7.2

1303 63.1 72.5 984 28.3 27.1 36.9 35.9 27.2 9.0 45.1 18.0 32.8 11.1

1261 66.3 69.4 942 – 27.1 36.9 36.3 30.0 9.0 45.1 21.2 34.7 10.2

863 50.6 41.0 611 26.0 30.0 39.8 42.1 8.5 8.5 47.0 5.6 5.8 2.0

907 45.8 51.2 655 27.3 25.8 35.6 35.4 10.4 8.5 42.8 5.0 8.8 4.7

VL – lipid molecular volume, D – lamellar repeat spacing (see Figure 3.23), A – average interfacial area/lipid, VC – sum of volumes of chain methylenes and methyls, VCH3 – volume of methyl, DC – thickness of hydrocarbon core (DC ¼ VC/A), DHH – head group peak–peak distance, DB – Gibbs–Luzzati bilayer thickness (DB ¼ 2VL/A), DW – Gibbs–Luzatti water thickness (DW ¼ D  DB), DH0 - steric head group thickness (DH0 ¼ (DB0 /2)  DC), DB0 – steric layer thickness (DB0 ¼ 2(DC þ DH0 ), DW0 – steric water thickness (DW0 ¼ D  DB0 ), nW – number of water molecules/lipid (nW ¼ ADW/2VW), nW0 – number of waters between DC and DB0 /2 (nW0 ¼ ADH  VH)/VW) (Reprinted from Biochim.

Biophys. Acta, 1469, Nagle and Tristram-Nagle, Structure of lipid bilayers, 37. Copyright 2000, with permission from Elsevier.).

Biological Membranes and Membrane Mimics

is expected that removal of the water from the interbilayer space, as a result of increased osmotic pressure, should squeeze membranes together and, in addition, should result in a decrease in the area per molecule (or an increase in the membrane thickness) [94]. The changes in the thickness are, however, rather small, 0.12 nm when osmotic pressure is increased from 0 to 5.6 · 106 N m2 [6]. Analysis shows that hydration and undulation forces, as well as van der Waals attractive forces, contribute to the total interbilayer pressure. The hydration pressure is dominant for interbilayer distances of 0.5–1.3 nm, the undulation pressure is greatest at larger distances and van der Waals forces at lower spacing. Interbilayer interactions depend on the kind of phospholipids involved. For charged lipids in low salt concentrations, one should also consider electrostatic interactions. Glycolipid bilayers are characterized by more complex interactions and also include strong adhesive forces due to the saccharide head groups [94]. 3.4.3 Dynamics and Order Parameters of Bilayers Determined by EPR and NMR Spectroscopy and by Optical Spectroscopy Methods Electron paramagnetic resonance (EPR), nuclear magnetic resonance (NMR) and optical spectroscopy methods allow one to obtain information about lipid chain configurations and dynamics and study the mechanisms of protein–lipid interactions. The theory of these methods is well described in the literature (see, for example, [95]). The radio spectroscopic methods (EPR, NMR) are based on the interactions of electron spins (EPR) or proton spins (NMR) with a magnetic field. If the spins are not paired, the spin of electrons or overall spin of the charged nucleus generate a magnetic dipole along the spin axis. The magnitudes of these dipoles are fundamental properties of electrons and nuclei and are called electron (me) and nuclear (mN) magnetic moments, respectively. If a single electron or a single proton is placed in a static magnetic field H, the energy of the spins will be split in to two energy levels. The energies of these levels are gbH/2 and gbH/2, respectively. Thus, the difference between the upper and lower energy levels is DW ¼ gbH, where g is the magnetogyric ratio and depends on the nature of the paramagnetic particle. For electrons g  2 and b ¼ 0.977 · 1023 J/ T is the Bohr magneton. The number of electrons that will occupy the lower energy level is N1 and that for the upper level N2. The ratio N1/N2 ¼ eDW/kT ¼ egbB/ kT > 1. Thus, the number of electrons at the lower energy level will be higher than at the upper one. If an alternating electromagnetic field with a frequency n is directed perpen-

111

dicular to the magnetic field H, the electrons from the lower energy level will be moved to the upper energy level, i.e. the system will adsorb energy. The maximum energy absorption will take place at resonance conditions. In the case of EPR, this resonance frequency is nre ¼ gbH=h

ð3:23Þ

where h is Planck’s constant. It is more convenient to keep frequency constant and change the magnetic field intensity, H. In this case the resonance energy absorption will take place at H re ¼ h n=gb

ð3:24Þ

The EPR spectra represent the dependence of the intensity of absorption on the magnetic field strength. Because phospholipid molecules are diamagnetic, they must be labeled by spin labels. Alternatively, spin probes can be used for incorporation into the lipid membranes in order to measure an EPR signal. Figure 3.24 shows the structural formulas of two typical spin probes with different locations of nitroxyl radical. It also shows the location of the probes in the membrane and the spectra of the probes. The EPR spectrum of quickly rotating I1,14 (1-doxylstearate) represents a triplet involving narrow components. Any slowing down of the label rotation or any movement anisotropy is associated with a visible change in the nitroxyl radical spectrum. The spectrum is typical for the spin label I12,3 (12-doxylstearat), the nitroxyl radical of which is located close to the polar head group region, which is more ordered than the inner hydrophobic region of the membrane. The EPR spectrum and its components are dependent on the anisotropy of the environment as well as on the ordering, micro viscosity, surface charge, etc. Thus, increased environmental anisotropy broadens spectral lines, which become more distant from each other. The parameter AII, distance between outer maxima (Figure 3.24), is sometimes employed as it depends on the label roation speed and on the degree of the label with orientation respect to the membrane surface. The relative amplitudes of the individual spectral components are also changed. Changes in environmental anisotropy are followed by changes in the rotational diffusion of the label. Quantitatively, rotational diffusion is characterized by the rotational correlation time tR tR 1  2pðAzz  Axx Þ=h

ð3:25Þ

where (Azz  Axx) is the maximum anisotropy of the 14 Nhyperfine splitting of the nitroxyl radical [96]. The

112

Bioelectrochemistry

Figure 3.24

(a) Structural formulae and location of the spin labels I1,14 and I12,3 in a membrane and (b) their EPR spectra.

characteristic rotation times for lipids in the fluid state are around 109 s. Characteristic times for lipids restricted by interactions with proteins are in the region tR  1–5 · 108 s [97]. The ordering parameter S is an informative characteristic of EPR spectra. This parameter characterizes the degree of ordering of the bilayer. The parameter S is determined by [99]



AII  A? Axx þ Ayy þ Azz  2Axx  ðAxx þ Ayy Þ=2 AII =2 þ A?

ð3:26Þ

Where Axx ¼ Ayy ¼ 0.58 mT and Azz ¼ 3.1 mT. The order parameter depends on the structural state and composition of the bilayer. In general the parameter S decreases toward the center of bilayer, which is consistent with increased mobility of the hydrocarbon chains toward the methyl groups (Figure 3.25) [99]. Despite a number of advantages of the EPR spin probe method, questions may arise about possible disturbance of the membrane by the spin probe. This problem does not

exist for NMR. This is due to the fact that many nuclei have their own magnetic moment, which is sensitive to their surroundings, i.e. they allow the measurement of an NMR spectra. The resonance condition for NMR is given by the relation H RN ¼ nh=gN mN

ð3:27Þ

where gN is the nucleus magnetogyric ratio and mN is the nuclear magneton. The magnetic moment is typical for nuclei 1 H, 15 N, 19 F, 31 P etc., but not for 4 He, 14 O, 12 C. In biological molecules there are many protons, which gives the possibility of applying this method to the investigation of the ordering and dynamics of biomembranes. The spectral lines depend on the chemical environment of the atoms. This is the so-called chemical shift. Unfortunately, a large numbers of chemically different protons have similar values of chemical shift. This problem can be solved by the selective deuteration of the molecules, e.g. lipids, using isotopes 2 H or 13 C. If hydrophobic chains of phospholipids are predeuterated, the order parameter S can be determined.

Biological Membranes and Membrane Mimics

Figure 3.25 The plot of the ordering parameter S on the position of the hydrocarbon n, for fully hydrated multilayers of A – eggPC and B – eggPC þ cholesterol (molar ratio 2 : 1). [99] (Reprinted from Spin Labelling, Theory and Applications, L.J. Berliner, H.M. McConnell, Molecular moving in biological membranes, Copyright Elsevier 1976.)

This parameter decreases towards the center of the bilayer as in the EPR method. The study of the structure and dynamics of molecules has been successfully performed using 31 P NMR. 31 P nuclei are often present in biological molecules. These spectra are sensitive to the structure of the lipid bilayer. This is due to the fact that lipid phosphorus exhibits a large chemical shift anisotropy. In large liquid crystalline systems, such as biomembranes (linear dimensions > 200 nm), the rotation of the lipid molecule along its long axis is only partially averaged. Therefore, instead of narrow component, the NMR spectrum is broadened and is composed of a low field shoulder and high field peak (Figure 3.26). This shape of the 31 P NMR spectra is typical for large multilayer systems, e.g. MLVs. In contrast to MLVs, in small sonicated unilamellar vesicles (ULV), the fast diffusion of the lipid molecules produce a line-narrowing effect. Narrow NMR spectra are also typical for micelles and cubic or rhombic lipid phases [100,101]. However, when lipids are in the HII hexagonal phase (Figure 3.26), additional motional averaging takes place

113

due to lateral diffusion around small (2 nm diameter) aqueous channels. This effect results in a characteristic 31 P NMR line shape with the reverse asymmetry to the bilayer spectra (Figure 3.26) [101]. Thus the 31 P NMR spectra can be used to study the formation of non-lamellar phases in lipid systems. The non-lamellar systems – hexagonal (HII) phases – can appear in initially lamellar phases of lipid bilayers composed of phosphatidylethanolamines at higher temperatures. The transport of ions can also induce the HII phase [102]. Non-lamellar phases can appear also around proteins in a membrane [103]. Analysis of NMR spectra reveals that, in the liquid crystalline state of the bilayer fast rotation, characterized by a relaxation time tR  107 s1, takes place. At the temperature of the phase transition of phospholipids from a less ordered liquid crystalline phase to a more ordered gel phase at lower temperature, the micro viscosity of the membrane increases. This is reflected by the increased time of molecule rotation tR > 105 s. However, the mobilities of polar head groups of phospholipids are less sensitive to the phase transition in comparison with hydrocarbon chains, and the relaxation time for the rotation of the head groups is less affected by temperature. The sensitivity of 31 P NMR to the structural state of the membrane makes this method a sensitive tool for the study of the phase transitions in lipid bilayers. For example, the half width DY1/2 of the sharp peak in the 31 P NMR spectra of MLVs is influenced by dipole–dipole interactions between phospholipid head groups. This interaction changes during the phase transition of the bilayer. Therefore, the phase transition can be studied by the measurement of the dependence of DY1/2 on the temperature [104]. Among optical spectroscopy methods, fluorescent spectroscopy is especially suitable for the study of the physical properties of lipid bilayers. Phospholipids do not show fluorescence, therefore they should to be labeled by fluorescence labels, or fluorescence probes should be used to monitor the membrane properties, because the probes are sensitive to their surroundings. Currently there exist a large variety of fluorescent probes that can be used for the study of various aspects of membrane biophysics and bioelectrochemistry. The amphiphilic probes sensitive to the membrane anisotropy (e.g. 1,6 diphenylhexatriene (DPH)) or potential-sensitive styryl dyes (e.g. RH 421 or di-8-ANEPS) can serve as examples. The application of fluorescent probes to study the ordering and dynamics of the membrane is based on the measurement of fluorescence intensities III and I? in two directions of polarization with respect to the direction of the exciting beam. The fluorescence anisotropy is characterized by the parameter r

114

Bioelectrochemistry

Figure 3.26 The schematic structure of some phospholipid phases and their

III  I ? r¼ I II þ 2I ?

r0  r/ þ r/ 1 þ t=tc

P NMR spectra (According to Ref. [100])

are connected with the order parameter S [105] ð3:28Þ

Thus, if the probe is in a fully isotropic environment, r ¼ 0. The membrane, however represent an anisotropic body, therefore r 6¼ 0 and the fluorescence anisotropy will depend on the structural state. The anisotropy measured in steady-state conditions can be represented as r¼

31

ð3:29Þ

where r0 and r/ are the initial and limiting values of timeresolved anisotropy, tR is the correlation time and t is the lifetime of the fluorescence probe. The parameters r0 and r/

S2 ¼ r / =r 0

ð3:30Þ

Fluorescence spectroscopy is useful for studying thermal phase transitions and the mechanisms of interaction of various species with lipid bilayers. As an example, on Figure 3.27 there is a plot of fluorescence anisotropy as a function of the temperature obtaine, with DPH dye in vesicles composed of DPPC, in complexes of DPPC with DNA from salmon sperm and with dextran sulphate in the presence of magnesium ions. For pure DPPC vesicles the parameter r changes sharply with increasing temperature in the region of the pretransition (T  35  C) and at the main transition temperature (T  41  C). For the complex DPPCþDNAþMgCl2 only, a monotonous decrease of r

Biological Membranes and Membrane Mimics

115

Figure 3.27 The temperature dependence of fluorescence anisotropy r for the fluorescence probe DPH in DPPC vesicles ( ), DPPC in a complex with DNA (&) and DPPC in a complex with dextrane sulphate (~). In the last two cases magnesium ions were also present in the electrolyte. All compounds were in equimolar concentrations, 2.5 · 104 mol l1 and the concentration of the fluorescence probe was 2 · 107 mol l3 [106]. (Reprinted from FEBS Letters, 137, Gruzdev, Khramtsov, Weiner and Budker, Fluorescence polarisation study of the interaction of biopolymers with liposomes, 4. Copyright 1982, with permission from Elsevier.)

Figure 3.28 Fluorescence excitation spectra of RH 421 in an aqueous solution of 0.2 mg ml1 DPPC vesicles as the control (–) and in the presence of 0.5 mM gramicidin A (....) or 10 mM phlorentin (----) [109]. (Reproduced by permission of Biophysical Journal.)

was obtained. It is possible that interaction of DNA with the head groups of DPPC results in partial denaturation of the DNA and as a result more hydrophobic bases interact with the hydrophobic part of the bilayer. The interaction of dextran sulphate (DS) with DPPC is rather strong as is evident from Figure 3.27. It can be seen that a second DS-induced phase transition at temperature higher than main transition of phospholipids is present. This effect may be due to the formation of hydrogen bonds between hydroxyl groups of DS and the carbonyl groups of DPPC [106]. Potential sensitive dyes, like RH 421 or di-8-ANEPS can be used as a sensitive tool to monitor changes in the membrane potential. They reveal the so-called electrochromic effect, i.e. the electric field causes a shift in their fluorescence spectra [107]. The styryl dyes, like RH 421, are amphiphilic with a partition coefficient Clipid/Cwater > 105 [108]. Due to the amphiphilic nature of the dye, its polar part is located in the region of the polar head groups of the phospholipids, while the chromophore and the hydrophobic tail of the molecule are located in the hydrophobic part of the membrane. The sensitivity of the dye to changes in the dipole potential of the membrane is illustrated in Figure 3.28,

where the normalized fluorescence excitation spectra of RH 421 in DPPC vesicles with and without gramicidin (5 mmol l1) and/or phloretin (10 m mol l1) are presented. The shift in the spectra shows that both species change the dipole potential of the DPPC bilayer [109]. Quantitative information about the changes of dipole potential can be obtained by the determination of the ratio of fluorescence intensities detected at two excitation wavelengths on the blue and red edges of the excitation spectra [107]. For RH 421 dye the ration is: R ¼ I440/I540, where the indices correspond to the wavelength of the excitation beam. For example, Shapovalov et al. [109] showed that gramicidin A produces a pronounced decrease in the R value with increasing concentration of this channel former. It was supposed that gramicidin reduces the existing positive dipole potential of bilayers by inducing reorientation of dipole-carrying groups (cholines or hydrated water molecules). Ion translocation could change the local electric field in the membrane [110], therefore the electrochromic effect of styryl dyes can be used to study the charge movement across membranes with incorporated ionic pumps [110]. If a membrane containing fluorescent dyes is briefly illuminated by an intens laser beam, the dyes in the light exposed area lose their fluorescence. The fluorescence of

116

Bioelectrochemistry

this part, however, starts to increase with time, due to diffusion of undamaged dyes from the surrounding membrane surface that was not illuminated by the laser. This method, called photo-bleaching, allows the determination of the coefficients for lateral diffusion of lipid bilayers (6 · 1012 m2 s1) or proteins (1014 m2 s1). A large variety of modern laser spectroscopic methods, such as Raman spectroscopy and time resolved nano- and picosecond spectroscopy are also of considerable interest in membrane studies [111].

3.5 PHASE TRANSITIONS OF LIPID BILAYERS Bilayer lipid membranes can exist in different phases depending on the water content and temperature. Transition between phases can be induced by varying either the lipid concentration (lyotropic mechanism) or the temperature (thermotropic mechanism). For biomembranes, of particular interest are the transitions involving the lamellar or bilayer lipid phase. The phase transition processes typical for biomembranes are connected with hydrocarbon chain-melting transitions and transitions to nonlamellar phases. The chain-melting transition is based on the configuration entropy of the hydrocarbon chains. The driving force for the transition from a lamellar to a non-lamellar phase is the tendency for spontaneous curvature of the bilayer phase. This process has great significance for cell or vesicle fusion [112]. The various phases have been classified by nomenclature proposed by Luzatti [113]. A Latin letter characterizes the type of long-range order: L – one-dimensional lamellar, H – two-dimensional hexagonal, P – two-dimensional oblique, Q – three-dimensional cubic, C – three-dimensional crystalline. A lower-case Greek subscript characterizes the short-range conformation of the hydrocarbon chains: a – disordered (fluid), b – ordered untilted (gel), b0 – ordered tilted (gel). A Roman numeral subscript is used to characterize the content of the structure element: I – paraffin in water (normal), II – water in paraffin (inverted).

3.5.1 Lyotropic and Thermotropic Transitions Lyotropic Transitions In general, lyotropic transitions between single phases will take place via two coexisting phases. The lyotropic mesomorphism found in lipid–water systems can be schematically represented by a diagram (Figure 3.29) [112,114]. At very low lipid concentrations, below the

Figure 3.29 Schematic lyotropic lipid–water phase diagram for phospholipids. This is the high temperature section of the temperature composition, and is intended primarily to indicate phase transitions induced by varying water content. Hatched areas indicate two-phase domains. The left-hand side of the diagram (from the La phase to higher water content) is representative of single-chain phospholipids (V/Al 1) [112]. (Reprinted from Chem. Phys. Lipids, 57 Marsh, General features of phospholipid phase transitions, 12. Copyright 1991, with permission from Elsevier.)

critical micelle concentration (cmc), the lipid is in the form of monomers. At concentrations higher then the cmc, the lipids form normal micelles (MI). With further increase in the lipid concentration, the system is transformed, first to a normal hexagonal phase (HI) (probably via an intermediate cubic phase Q0 I). HI is then transformed to the normal cubic phase (QI) and then to the lamellar phase (La). This sequence is typical for singlechain phospholipids. For two-chain phospholipids at higher concentrations the lamellar phase is transformed into an inverted cubic phase (QII) and then to an inverted hexagonal phase (HII).

Thermotropic Transitions Because in excess water the composition is fixed, a single phase can exist over a range of temperatures. However, two phases can coexist only at a fixed temperature. Therefore, sharp thermotropic phase transitions occur at certain temperatures. In general, the sequence of thermotropic transitions of hydrated phospholipids can be represented by the following scheme

Biological Membranes and Membrane Mimics

TS

TP

Tt

Th

Tl

LC ! Lb ! Pb0 ! La ! QII ! H II ! M II

ð3:31Þ

With increasing temperature, the sub-transition from the crystalline phase (Lc) to the hydrated lamellar gel phase (Lb) takes place at Ts. The lipid chains can be tilted or not tilted with respect to the bilayer normal. Then, at a temperature Tp, the pretransition from a low-temperature gel phase to an intermediate ripple phase (Pb0 ) takes place. The main transition from gel to fluid lamellar phase (La) occurs at temperature Tt. The phase La is characterized by disordered lipid chains. With increasing temperature, the fluid phase undergoes further transitions, first to an inverted cubic phase (QII) and then to an inverted hexagonal phase (HII). The final stage is the transition into the inverted micellar phase (MII) that occurs at temperature Tl. This phase is characterized as an immiscible oil in excess water. Not all the phases and transitions mentioned above appear for a single phospholipid [112].

3.5.2 Thermodynamics of Phase Transitions At the phase transition temperature, the structure and properties of the lipid bilayer change sharply within a narrow temperature interval – usually less then 0.1  C. The changes in the ordering of the lipid bilayer are connected with changes in the entropy of the system. In addition there are sharp changes in the volume of the phospholipids. This is a typical feature of the first-order transition. The change in Gibbs energy at the transition temperature (Tt) is zero, i.e. DG ¼ DHt  TtDSt ¼ 0 and thus the changes in the entropy of the system at the phase transition temperature are connected with changes in the system enthalpy DSt ¼ DH t =T t

ð3:32Þ

For a first-order transition, the change in transition temperature with pressure P should be directly related to the change in volume DVt at the transition, via the Clausius–Clapeyron equation dT t =dP ¼ N A DV=DSt

dT t =dg ¼ 2N A DA=DSt

117

ð3:34Þ

where the factor of two allows for the two halves of the bilayer. This equation has been verified in [116]. Various effects, particularly connected with the composition of the aqueous phase, e.g. ionic strength, can induce a shift in phase transition temperature (see [112]). For charged lipids, shifts in transition temperature arise from the difference in surface charge density in the two phases. This shift can be estimated using electrostatic double layer theory (see below): DT elt ¼  ðe=pÞðkT=eÞ2 N A kf½1 þ ðs=cÞ2 1=2  1gDAt =DSt ð3:35Þ where k ¼ (8pNAe2I/1000ekT)1/2, is the reciprocal Debye screening length and c ¼ (ek/2p)(kT/e). The electrostatic shift is determined mostly by the charge density s, the ionic strength I and by the change in the area/molecule, DAt. The experimental determination of the variation of the phase transition temperature due to electric effects has been performed by Tr€auble et al. [117]. Experimentally the phase transition can be determined by measuring the transition enthalpy by scanning calorimetry. The calorimetric properties can be presented as a specific heat capacity or changes in the enthalpy as a function of temperature. An example of the calorimetric properties of large unilamellar vesicles of DMPC is shown in Figure 3.30, together with a plot of the specific volume as a function of temperature. The specific volume, V ¼ [1  (r  r0)/c]/r0 of the phospholipid can be determined on the basis of precise measurement of the density r using, for example, the vibrating tube principle [118]. Here r is the density of the lipid solution, r0 is the density of the buffer and c is the concentration of lipids. The low temperature peak in heat capacity at T ¼ 15  C is connected with the pretransition, i.e. the transition from ripple gel phase to lamellar gel phase. The narrow peak at Tt ¼ 24.1  C is connected with the main transition from gel to fluid lamellar phase. From the changes in specific volume we can see sharp changes at the main transition temperature confirming the first-order nature of the main transition [119].

ð3:33Þ

where DSt is the transition entropy and NA is Avogadro’s number. This relation is valid for phospholipid bilayers [115]. Similarly, the change in transition temperature in response to an isotropic membrane tension g, is related to the changes in the bilayer area DAt at the phase transition

3.5.3 Trans–Gauche Isomerization Microscopically chain melting is connected with rotation around the carbon bonds of the phospholipid hydrocarbon chains. The lowest energy holds for trans and highest for

118

Bioelectrochemistry

Figure 3.30 Calorimetric and specific volume properties as a function of temperature for large unilamellar vesicles of DMPC: (a) excess heat capacity; (b) specific volume.

cis conformations of the chains. In the gel state the rotation is restricted and the saturated chains are in the trans conformation. When the temperature approaches the phase transition region, the probability of rotation increases. Rotation by 120 relative to the trans conformation results in the formation of gauche (þ) or gauche () conformations. Energetically the gauche conformation does not differ substantially from the trans conformation (2–3 kJ mol1), however these two conformations are separated by a relatively high energetic barrier (12–17 kJ mol1). The appearance of the gauche (þ) conformation causes steric difficulties in a bilayer. However, subsequent gauche () rotation results in a diminishing sterical repulsion. As a result of the gauche (þ)–gauche () rotation, a kink conformation appears in the lipid chain (Figure 3.31). In this case the spatial configura-

tion of the chain is preserved, but the chain is shorter by 0.127 nm and the cross-sectional area increases. The phase transition in a lipid bilayer from a gel to a liquid state is therefore accompanied by a decrease in the thickness and an increase in the area per molecule. The volume of phospholipid changes to a lesser extent. The presence of unsaturated phospholipids considerable increases the probability of trans–gauche isomerization and therefore the phase transition temperature decreases. In order to estimate the effect of trans–gauche isomerization, let us compare the frequency of torsional oscillations in CC bonds (7 · 1012 s1) with the frequency of appearance of the gauche conformation at room temperature (300 K). Considering that the energetic barrier separating the trans and gauche conformations is DE ¼ 12 kJ mol1 we have: n ¼ (kT/h) exp [DE/(RT)]  1010 s1 [61]. Thus, the gauche conformation appears with a high frequency due to torsional oscillations. In the fluid state the kink can move along the chain due to synchronous rotation by 120 of the corresponding CC bond. The shift to the neighboring position is of the order of DL ¼ 0.13 nm. The shift of the kink can be considered as one-dimensional diffusion along the chain. This diffusion can be characterized by a diffusion coefficient Dk ¼ 0.5n (DL)2  105 cm2 s1 (here we assume that the frequency of the jump of the kink is of the order of 1010 s1). This value practically coincides with the diffusion of oxygen, water or small molecules of non-electrolytes through a lipid bilayer. Thus it can be assumed that the transport of some species through the lipid bilayer can be due to the appearance of free volume in a membrane, formed by kinks.

3.5.4 Order Parameter Figure 3.31 The conformation of the hydrocarbon chain of a phospholipid. A – trans configuration, B – gauche-trans-gauche conformation, C – cis-trans-gauche configuration.

The high mobility of hydrocarbon chains allows us to determine only the average, most probable orientation of

Biological Membranes and Membrane Mimics

the chains. Even in a gel state there exists conformational mobility of the chains that increases toward the center of the bilayer. In order to describe the shift in the orientation of the chain from the direction normal to the bilayer surface, the order parameter Sn is usually used 3 1 Sn ¼ cos 2 Qn  2 2

ð3:36Þ

where Yn is the angle between the normal of the bilayer and the normal to the plane formed by two vectors of the CH bonds of the nth segment of the hydrocarbon chain. Obviously, for an ideally ordered chain, S ¼ 1, and for disordered phase, S ¼ 0. The order parameter of the bilayer is almost constant up to 8–9 methyl segments, but decreases substantially after this segment. This has been established by both NMR and EPR spectroscopy [6,97]. It is possible, that the initial segments of the hydrocarbon chain provide cohesive interaction between the chains, that is, in addition to the hydrophobic interaction, necessary for preserving the bilayer integrity. It is interesting that in most of the natural phospholipids the double bonds in unsaturated fatty acids start after the 9th carbon atom, and thus do not decrease the ordering of the densely packed initial parts of the chains. 3.5.5 Cooperativity of Transition The chain-melting transition of phospholipids with saturated hydrocarbon chains is highly cooperative, with a transition width that can be less than 0.1  C. Phenomenological theories of the cooperativity of the phase transition have been formulated on the basis of the coexistence of clusters of lipid molecules. The cooperativity of the transition is evaluated as the ratio between the Van’t Hoff enthalpy and the calorimetric enthalpy: DHVH/DHCAL pffiffiffiffiffiffi ¼ 1/ s0 , where s0 is the parameter of cooperativity and the pffiffiffiffiffiffi value 1/ s0 is the size of the cooperativity unit (s0 ¼ 1 corresponds to non-cooperative behavior, while a value of s0 < 1 represents cooperative behavior; the lower the value of s0, the higher the cooperativity of the system). The Van’t Hoff enthalpy can be approximated from the half width of the transition DT1/2: DHVH ffi 7T 2t /DT1/2 [61]. For example, the number of lipid molecules in cooperative units for DPPC was estimated as 70 10 and that for DMPC 200 40. Due to thermal fluctuations the liquid phase is created in a gel phase. At the phase transition temperature both gel and liquid phases coexist. With increasing temperature the number of molecules in a gel phase dramatically decreases. Fluctuations in cluster size take place throughout the transition [120].

119

3.5.6 Theory of Phase Transitions The simplest theory of phase transitions in lipid bilayers has been proposed by Nagle [121] and is based on the ordering–disordering transition, assuming the existence of trans–gauche conformations in each hydrocarbon chain (see above). The temperature of the phase transition, calculated on the basis of this theory, was close to that obtained by calorimetry. In the model by Marcelja [122], the energy of a chain with configuration i in a bilayer is determined as follows E i ¼ E iin þ E idisp þ pAi

ð3:37Þ

where E iin is the energy of the chain connected with the trans–gauche transition. The second term E idis is connected with the intermolecular disperse interactions and the third term (pAi) is connected with the existence of lateral pressure in the bilayer due to steric repulsions, electrostatic interactions and the hydrophobic effect. The model of Marcelja allows us to calculate the phase transition temperature and the enthalpy of the transition, as well as the basic geometrical parameters of the chains all in good agreement with experiments. The phase transitions in a bilayer can also be described in general by the phenomenological theory by Landau [123]. This theory allows us to calculate free energy near the phase transition temperature

Gh ¼ a 1 h þ

1 1 1 a2 h2  a3 h3 þ a4 h4 2 3 4

ð3:38Þ

where Z is the ordering parameter, a1 ¼ p(Af  Ag), p is the lateral pressure, Af and Ag are the areas per lipid molecule in the liquid and gel states, respectively. Coefficients a2, a3 and a4 can be found from the dependence of Tt and Z on the lateral pressure (p). The parameters of the phase transition can be found from the minima of the function GZ (T, Z). With the theory of Landau, the ordering parameter is determined through the area per molecule at the transition temperature h ¼ ðAf  AÞ=ðAf  Ag Þ

ð3:39Þ

where A is the real area per molecule in a bilayer. Using the Landau theory it has been possible to estimate the influence of cholesterol and proteins on the phase transition temperature. The results of this work are in agreement with the theory developed by Marcelja.

120

Bioelectrochemistry

The microscopic Pink lattice model [124] is based on the description of the conformational properties of an acyl chain by a small number of conformational states. The conformational chain variables are coupled by hydrophobic anisotropic van der Waals interactions. The interaction between the hydrophilic moieties is modeled by a Coulomb-type force or simply by an effective intrinsic lateral pressure. A detailed description of the phase transitions in lipid bilayers is given in a book by Cevc and Marsh [3].

3.6 MECHANICAL PROPERTIES OF LIPID BILAYERS 3.6.1 Anisotropy of Mechanical Properties of Lipid Bilayers Viscoelastic properties have a significant role in allowing biomembranes to perform different functions. Together with the cytoskeleton, viscoelasticity determines the cell shape and transduction of mechanical deformation from mechanoreceptors to sensitive centers. In addition, during conformational changes of the proteins the physical properties of the membrane can also change. These changes can be described by means of a macroscopic approach using the theory of elasticity of solid bodies and liquid crystals. The structure of the lipid bilayer is considerably simpler than that of a biomembrane. However, even the structure of the BLM has clearly expressed anisotropy. This leads to strong anisotropy in membrane viscoelastic properties and requires the description of bilayer properties by several elasticity moduli. The difficulties in describing membrane elasticity are not exhausted by this phenomenon. The behavior of deformable solid bodies is described by the theory of elasticity. The principal differences between biomembranes and the classical objects of this theory are as follows. One of the membrane properties, its thickness, is very small and is only 20–200 atoms in size. Therefore the influence of the microheterogeneity of each atomic layer on the membrane properties can be substantial. In this case the macroscopic parameters of the membrane, which are the result of the average of their properties over the environment, can considerably differ from the corresponding parameters of the membrane at the level of some distinguished layers. On the other hand, 20–50 atomic layers is too large a value to enable their description by equations of the theory of elasticity for each layer. Therefore there exist a number of models of biomembranes as elastic bodies, which average the properties of biomembranes over a certain number of such layers [125–129]. However, for subsequent analysis it is necessary to introduce a macro-

Figure 3.32 Schematic representation of membrane deformation. Arrows indicate the application of mechanical stress. For explanation, see the text.

scopic description of the membrane as an elastic body and to generalize it to account for its viscous properties. This analysis has been performed in a monograph by Hianik and Passechnik [51]. It has been shown that the understanding of the membrane as a viscoelastic body requires analysis of membrane deformation in different ways (Figure 3.32): (1) volume compressibility; (2) area compressibility; (3) unilateral extension along the membrane plane; (4) transverse compression. The mechanical parameters that characterize the membrane deformability listed above are the volume compressibility modulus K and the Youngs moduli of elasticity EII, E10 and E?, respectively. These parameters are defined as follows K ¼  p=ðDV=VÞ; E10 ¼ sx =U xx

EII ¼ sx =ðDA=AÞ ¼ 2sx =ðDC=CÞ E? ¼  p=U zz ¼ 2p=ðDC=CÞ ð3:40Þ

where sx is the mechanical stress along the membrane plane, p is the pressure compressing the membrane, Uzz and Uxx are the relative membrane deformations in the transverse direction and along the membrane plane, respectively, and DV/V, DA/A, DC/C are the relative changes in the volume, area and electrical capacitance, respectively. Owing to the small dimensions of the membrane, special methods were developed for measurement of the elasticity moduli. Below we briefly describe the basic methods for the measurement of the elasticity moduli of lipid bilayers and show their typical properties. The mechanical properties of lipid bilayers have been described in detail elsewhere [51]. Transverse Elasticity Modulus E? The transverse deformation, i.e. the parameter Uzz ¼ Dd/d (d is the membrane thickness) cannot be measured directly

Biological Membranes and Membrane Mimics

due to the small thickness of the membrane and the extremely small changes in the thickness upon deformation. Therefore the transverse deformation is determined mostly from measurement of changes in the electrical capacitance of the membrane. In the case of isovoluminous deformation, i.e. when the volume compressibility K is much higher then E? (this has been clearly shown experimentally [51]): Dd/d ¼ DC/2C (i.e. the decrease in thickness results in an increase in the membrane capacitance). In the transverse direction, the membrane cannot be deformed by mechanical pressure. However, because the membrane behaves electrically as a capacitor, when a voltage is applied to the BLM, it will compress the membrane with an electrostriction pressure p ¼ CSU2/2d (CS is the specific capacitance of the membrane CS ¼ C/A and U is the applied voltage). Therefore E? ¼ p/(Dd/d) ¼ 2p/(DC/C). To measure the changes in capacitance a special method is also required. This is connected with the inhomogeneity of the membrane and with the presence of a thick Plateau–Gibbs border. For example, if a dc voltage is applied to the BLM and the capacitance is measured e.g. by a capacitance meter, then the measured changes DC/C will be not only due to the changes in the thickness, but also other factors, such as rebuilding of new bilayer parts from the Plateau–Gibbs border. As a result the determined elasticity modulus will be underestimated in comparison with the real value. This particularly explains the underestimated values of transverse elasticity moduli in earlier work (see [51] for a review). Therefore, a special electrostriction method based on the measurement of the amplitude of the higher current harmonics has been developed [130]. This method, as well as its application to various BLM systems, is described in detail in [51]. Briefly, if an alternating voltage of amplitude U is applied to the BLM via electrodes (e.g. Ag/AgCl electrodes), due to the non-linear dependence of the capacitance on the voltage (C ¼ C0(1 þ aU2), where C0 is the capacitance at U ¼ 0 and a is the electrostriction coefficient), higher current harmonics, with frequencies 2f, 3f, etc. and amplitudes I2, I3, etc., respectively, will be generated in addition to the basic first current harmonic (frequency f) of amplitude I1. The measurement of these amplitudes allows us to determine various parameters, particularly the absolute value of the elasticity modulus E ? ¼ 3C s U 20 I 1 =ð4dI 3 Þ

ð3:41Þ

where Cs is the specific capacitance of the membrane. If, in addition to the amplitude, the phase shift j between the first and the third current harmonics is also measured,

121

then the coefficient of dynamic viscosity Z can be determined h ¼ E? sinj=ð2pf Þ

ð3:42Þ

(see [51] for a detailed description of the method and experimental set up). Using this method it has been shown that the elasticity modulus and the dynamic viscosity depend on the frequency of the deformation and increase with the frequency. The value of E? also depends on the content of the hydrocarbon solvent in the BLM and increases with decreasing concentration of the solvent, reaching a value of 3.6 · 108 Pa for solvent-free membranes composed of glycerolmonooleate at a frequency of deformation of 1.3 kHz. This modulus is also extremely sensitive to the lipid composition and the content of cholesterol or other sterols. For example, the E? value increases with increasing length of the phospholipid hydrocarbon chains which indicates higher order in the hydrophobic part of the membrane due to more extensive hydrophobic interaction between the phosholipid chains [51]. On the other hand, E? decreases with increasing degree of unsaturation of the fatty acids, showing a decrease in the membrane ordering [131]. This elasticity modulus changes considerably upon interaction with a BLM of low molecular compounds, e.g. local anesthetics, or macromolecules, e.g. integral or peripheral proteins [51]. The method of measurement of E? has also been applied to supported lipid membranes and gives the possible study of affinity interactions [132] or the interaction with a BLM of nucleic acids and their complexes, with cationic surfactants [133]. Recently, excitation FTIR spectroscopy has been applied to study the electrostriction of supported lipid multilayers composed of DMPC [134]. A periodic rectangular electric potential induced periodic variation of the tilt angle of the hydrocarbon chains by 0.09 0.015 . This corresponds to a variation of the bilayer thickness of Dd ¼ 5.4 · 103 nm. The calculated Young’s elastic modulus, E? ¼ 1.8 · 106 Pa, was in good agreement with data obtained by the electrostriction method [132]. The electrostriction induced changes of the tilt angle of acyl chains of DMPC were also studied in detail by Lipkowski and coworkers [135]. The Area Expansion Modulus EII EII can be determined by the method of micropipette pressurization of giant bilayer vesicles [116], or by determination of the changes in electrical capacitance during periodic deformation of spherical BLMs [125]. The value of

122

Bioelectrochemistry

EII can only be measured over a limited range of frequencies (5–10 Hz). The typical values of EII for BLMs with a hydrocarbon solvent were in the range 107–108 Pa, which is more then 10 times higher than the values of E? for a similar BLM compositions at the lowest frequency of deformation (20 Hz). The area expansion modulus is less sensitive to the lipid composition and does not significantly depend on the length of the phospholipid hydrocarbon chains and their degree of unsaturation [136].

Experiments to determine various elasticity moduli revealed that these values can only be estimated over a limited range of frequencies: EII: 5–10 Hz, E10: 2–300 Hz, E?: 20 Hz–15 kHz and K : 7 MHz [51]. However, these values can be approximated in the frequency range 10–200 Hz [51]. It has been shown that the following inequalities hold for these elasticity moduli: E10, E?  < EII  K. Thus the BLM represents an anisotropic viscoelastic body. The corresponding model of BLM deformation should fulfill the above unequalities.

The Elasticity Modulus E10 E10 has been measured for the longitudinal distension of a cylindrical BLM formed between two circles, one oscillating and the other attached to an ergometer. Values of E10 106 Pa have been obtained for membranes of various compositions. They were independent of frequency over an interval of 30–200 Hz, i.e. they are determined by bilayer elasticity rather than viscosity [51]. Modulus of Volume Compressibility K K has been measured by determination of sound velocity in a suspension of small unilamellar liposomes. Values of K¼ (1.70 0.17) · 109 Pa have been determined by this method for liposomes composed of egg phosphatidylcholine. Values of a similar order have also been obtained for large unilamellar liposomes composed of polyunsaturated fatty acids [131]. Using the measurement of the elasticity modulus K, the mechanical and thermodynamic properties of liposomes of various compositions containing cholesterol [137], or modified by proteins [138], can be studied.

3.6.2 The Model of an Elastic Bilayer It has been shown [51] that the mechanical properties of a BLM cannot be described by the isotropic mechanical models by Wobshall [125] and by Evans and Skalak [139]. The recently discussed brush model of membrane mechanics composed of two isotropic layers well describes the behavior of area expansion and bending elasticity moduli [136]. However, the model does not provide information about the distribution of chains across the bilayer and thus does not consider the anisotropy of mechanical properties in the transverse direction. A three-layer model of deformation has been assumed to describe BLM anisotropy by Passechnik [129]. The two outer layers (thickness h1) have a modulus of elasticity E(1) and the inner layer (thickness h2) has a modulus of elasticity E(2)  E(1), like a sandwich (Figure 3.33). Mechanical stress in the membrane plane (measurement of EII) deforms the layers with a large elastic modulus (E(1)) and stress perpendicular to the membrane plane (measurement of E?)

Figure 3.33 Inhomogeneous mechanical models of a BLMs and the nature of membrane deformation upon transverse compression. The dashed line shows the position of the membrane parts following compression. (a) Scheme of a three-layer BLM structure on crosssection. (b) Sandwich with completely adhering layers. (c) Sandwich with free gliding of layers; (d) sandwich with microinhomogeneities and adhering layers. A – bilayer, B – microinhomogeneity with diameter rn. (e) and (f) Schemes of planar and rough BLM, respectively [51].

Biological Membranes and Membrane Mimics

deforms the ‘soft’ layer modulus (E(2)). Therefore one can expect that E?  EII. Deformation of the ‘sandwich’ depends on the degree of adhesion of the layers. From the analysis performed in [51], it follows that the three-layer model with different degrees of layer adherence (i.e. the sandwich with fully adhered layers, Figure 3.33b, and the sandwich with free gliding layers, Figure 3.33c) can only describe the properties of small parts of a BLM. To describe the deformation of all BLMs these parts must be separated by regions into which the ‘superfluous’ matter of the bilayer, which is squeezed out with the transverse compression of the BLM, will be adsorbed. For this purpose microinhomogeneities with thickneses not surpassing that of the BLM were included in the model [129]. As in models a–c (Figure 3.33), the model with inhomogeneities includes an inner layer, the elasticity of which is considerably less than that of the external ones. Obviously, the microinhomogeneities represent metastable formations originating at the moment of membrane formation, due to the fact that the solvent is unable to leave the BLM volume quickly and must be located somewhere. The cross-sectional size of the microinhomogeneities must be comparable with the membrane thickness d. In this case the microinhomogeneities do not contribute to the electrical capacitance of the BLM: C  d1, i.e. as without microinhomogeneities (see [51]). The hypothesis of microinhomogeneities allows us to explain why transverse compression of a BLM by an electrical field leads to changes in membrane capacitance. In this case mainly the inner ‘soft’ layer is deformed. Moreover, deformation is isovoluminous [125] and this leads to a bulging of the matter of the inner layer from the planar parts A to the microinhomogeneities B (dashed line in Figure 3.33d), so that it becomes ‘invisible’ (these parts do not contribute to the capacitance). For the planar bilayer (Figure 3.33c) the increase in capacitance in one place is compensated by its decrease in another place. Thus, the mechanical properties of a BLM can be qualitatively described by a three-layer elastic model composed of anisotropic elements with defects. More detailed analysis of the three-layer model of BLM elasticity is given in [51]. 3.6.3 Mechanical Properties of Lipid Bilayers and Protein–Lipid Interactions Protein–lipid interactions play an essential role in the functioning of biomembranes [140]. The specificity of these interactions is, however, under discussion. The exact lipid composition is probably not essential. For example, changes in the fatty acid chain composition caused by diet have no injurious effect on cell function. However, dietary-induced changes in lipid composition

123

are limited: some features of the fatty acyl chain composition are maintained constant, like the chain length between, typically, C16 and C20, with about half the chains being saturated and half unsaturated. This means that overall features of lipid composition, such as the length and saturation of the fatty acids are likely to be important for the membrane properties [141]. The length and saturation of the fatty acids are responsible for the creation of a certain thickness of membrane and its physical state, which are important factors in determining the protein–lipid interactions [140,141]. In certain cases the structure of the polar part of the phospholipids also plays a role in protein–lipid interactions. There is, however, evidence that a small number of special lipids are important for the function of the protein. For example, for Ca2þ ATPase it is phosphatidylinositol-4-phosphate. The binding of this lipid results in a two-fold increase in the activity of the calcium pump [141]. Specific activity with cardiolipin has been observed for cytochrome c oxidase [97]. Functioning of membrane proteins accompanied by changes in their conformation could influence the structure and physical properties of the surrounding lipid environment. The interaction of proteins with membranes includes both electrostatic forces (mainly peripheral proteins) and hydrophobic interactions (integral proteins). The structural and dynamic aspects of protein–lipid interactions can be investigated directly by various physical methods. The EPR spectra reveal a reduction in mobility of the spin-labeled lipid chains on binding of peripheral proteins to negatively charged lipid bilayers. Integral proteins induce a more direct motional restriction of the spin-labeled lipid chains, allowing the stochiometry and specificity of the interaction, and the lipid exchange rate at the protein interface, to be determined from EPR spectra. In this way a population of very slowly exchanging cardiolipin associated with the mitochondrial ADP–ATP carrier has been identified (see [142] for a review). Fluorescence spectroscopy is also effective for the study of protein–lipid interactions. In particular, Rehorek et al. [143] showed that, as a result of conformational changes of the integral protein, bacteriorhodopsin, the ordering of the lipid bilayer increases and a transmission of conformational energy occurs over a distance of more than 4.5 nm. The mechanical properties of the membranes are also very sensitive to conformational changes in lipid bilayers. The influence of bacteriorhodopsin on the structural state of spacious regions of planar bilayer lipid membranes (BLMs) was shown by means of measurement of the elasticity modulus E? [144]. It was shown that the area of lipid bilayer with an altered structure for each cluster of three

124

Bioelectrochemistry

bacteriorhodopsin molecules exceeds 2800 nm2. Moreover, as a result of the illumination of the BLM modified by bacteriorhodopsin, a considerable increase in E? occurred (more then five times) with further saturation to a stable level. This condition was preserved for several hours after the illumination was switched off, which shows that there is a possibility of mechanical energy accumulation in a membrane. For analysis of the mechanism of the protein–lipid interaction, the thermodynamic and mechanical properties of lipid bilayers and proteoliposomes are important. Owing to the possible different geometry of the hydrophobic moiety of proteins and that of lipids, as well as to the action of electrostatic and elastic forces, regions of altered structure may arise around protein molecules [51,145]. The formation of similar regions may be one of the reasons for the occurrence of longdistance interactions in membranes. It is very likely that hydrophobic interactions play a key role in the establishment of links between integral proteins and lipids. The rigid hydrophobic parts of membrane-spanning proteins cause a deformation of the hydrophobic lipid chains due to length matching. This leads to the stretching or compression of the hydrocarbon lipid chains, depending on the relation of the hydrophobic part of the proteins and the surrounding lipids [146]. Distortion of the membrane by proteins may cause lipid-mediated attractive or repulsive forces between proteins. The possible situations are presented in Figure 3.34. Due to changes in the ordering of the lipid bilayer, an increase or decrease of phase transition temperature, as well as changes in the membrane mechanical properties take place. Due to considerable problems with the isolation and purification of integral proteins and with the determination of their structure, only a few proteins have been analyzed so far in respect of their influence on the thermodynamic and mechanical properties of lipid bilayers. Using the mattress model of Mouritsen and Bloom [146] as well as the Landau–de Gennes theory of the elasticity of liquid crystals, it was possible to explain satisfactorily the changes in the phase transition temperature of proteoliposomes containing membranebound reaction center (RC) and antenna protein (LHCP) [148,149]. The role of elastic forces was, however, studied only in a small number of studies and the mechanism of its action is not clear yet. In this section we will briefly report the results of analysis of the mechanisms of protein–lipid interaction based on knowledge of the thermodynamic and mechanical properties of lipid bilayers with incorporated bacteriorhodopsin. De-

Figure 3.34 Schematic representation of liquid-mediated protein–protein interactions induced by a hydrophobic mismatch: (a) a matched membrane; (b, c) hydrophobic mismatch of the same sense results in lipid-mediated attractive forces; (d) a case of hydrophobic mismatch of the opposite sense resulting in repulsive forces between proteins (According to Ref. [147]).

tailed consideration of the theory of the mechanisms of protein–lipid interactions is given in [51,140]. As we mentioned above, the incorporation of a protein into a lipid bilayer leads to a distorted region of the membrane. This leads to changes in the phase transition temperature DT, which is a function of the protein concentration and according to [148] can be determined by the expression DT ¼ 8x2 ð2r 0 =x þ 1Þ½2ðd f  d p Þ=ðd f  d g Þ  1 xp

ð3:42Þ

where x is the characteristic decay length, r0 is the radius of the bacteriorhodopsin (BR) molecule, df and dg are the lengths of the phospholipid hydrocarbon chains in the fluid or gel states, respectively, dp is the length of the hydrophobic moiety of BR and xp is the molar ratio of BR and phospholipid (number of BR molecules/number of phospholipid molecules). The parameter x(T) is not measurable in the experiment. To determine its quantity, as well as

Biological Membranes and Membrane Mimics

125

perturbation UðrÞ ¼ u0 exp½  ðr  r0 Þ=x

Figure 3.35 (a) Schematic cross-section of an integral protein in a phospholipid membrane: a is the half-bilayer thickness and r0 is the radius of the protein; (b) components of membrane distortion that contribute to the free energy (Adapted from Ref. [151], reproduced by permission of the Biophysical Journal).

determining the energy of the elastic membrane deformation around the protein, we have used the algorithm described in [150,151] for the numerical calculation of the mechanical energy of the membrane around the ionic channel. Figure 3.35 shows a schematic cross-section of an integral protein in a membrane. The free energy change per unit area in cylindrical polar coordinates is

F ¼ 2p

Z

ð3:45Þ

In the calculations, elastic parameters and the thickness of the hydrophobic part typical for DMPC bilayers in the g gel (g) and fluid (f) states have been used: E ? ¼ 2 f 9 9 ˚ ; E ¼ 3:76 · 10 dyn  A ˚ 2 7:28 · 10  dyn  A ? (from measurements of solvent-free BLM [153]), gd ˚ 1, gf ¼ 3 · 108 dyn A ˚ 1 (measure15 · 108 dyn A 6 ments on liposomes [154]), K1 ¼ 10 dyn (a typical value for smectic mesophases [155]). The parameters that characterize the BR molecule are: dp ffi 30 Å and r0 ¼ 17.5 Å [156]. The thickness of the hydrophobic part of the lipid bilayer in gel and liquid crystalline states were as follows: dg ¼ 34.2 Å and df ¼ 22.8 Å, respectively (see [86] for the method of calculation). The results of the calculations of deformation energy, characteristic decay lengths x and x0 (see below) and the range of the distortion region r  r0 in the gel and fluid states of lipid bilayers of DMPC containing BR are shown in Table 3.7. The dependence of u(r), which represents the profile of the distorted region of the membrane around the protein in the fluid state, is shown in Figure 3.36. The exponential shape of the deformation (curve A) obtained using Equation (3.45) and the value x ¼ 16.8 Å (determined from the minima of free energy of the system) have a considerably larger range then those determined from the

rdr½E? u2 =a þ aK 1 ðu0 =r þ u00 Þ2 þ gðu0 Þ2

ð3:43Þ

where E?, K1 and g are the elasticity moduli of transverse compression, splay and surface tension, respectively. To determine the minimum energy conformation, we minimize the free energy with respect to the variation in u(x,y) and get the linear differential equation [151] K 1 ðu0 =r3  u00 =r2 þ 2u000 =r þ uIV Þ 00  ðg=aÞðu0 =r þ u Þ þ ðE ? =a2 Þu ¼ 0

ð3:44Þ

Equation (3.44) can be solved numerically using the algorithm described by Pereyra [152]. The parameter x can be determined from the minima of the free energy of the system with the assumption of exponential decay of

Figure 3.36 Geometry of the distortion region of a lipid bilayer around BR. (A) u(r) calculated according to Equation (3.43) from the minima F ¼ F(x); (B) u(r) from the exact solution of Equation (3.44); C backward transformation of exact solution to the exponential function (see the text) [51].

126

Bioelectrochemistry

exact solution of Equation (3.44) (curve B). Interestingly, the second curve is not exactly exponential in both the gel and fluid states. Using the backward transformation of the exact solution to the exponential function we obtain a new value of x0 ¼ 9.6 Å. The range of deformation is, in this case (curve C), about half as large in comparison with the exact solution (curve B). The remarkable region in Figure 3.36 is the section where all three curves cross. One can assume that this point determines the minimal distance from BR, from which the differences between the exact solution of u ¼ u(r) and the assumed exponential decay of the perturbation originate. The region between the BR surface and the point of differentiation can be considered as the region of immobilized influence of the protein on its lipid environment. Parameters for distorted regions (See Table 3.7) allow us to calculate the changes in the phase transition temperature with BR concentration xp, by means of Equation (3.42). In this calculation we used a similar method to Peschke et al. [148], which considers the aggregation of LHCP. The number of BR monomers in purple membrane clusters is known from RTG analysis (N ¼ 3) [157]. Therefore we modified Equation (3.42) for all possible combinations of parameters, and created dependencies of Tc on xp (Figure 3.37). In the case of x0 ¼ 10.48 Å, and assuming the aggregation of BR to trimers, we obtained a surprisingly good agreement with experiment. Similar results were also obtained for the gel state of the membrane. The values obtained show that, at the phase transition from the gel to the fluid state an increase in the deformation energy of the system BR-DMPC takes place (see Table 3.7). This means that the ordering of hydrocarbon chains of lipids around proteins in the fluid state increases, i.e. protein stabilizes phospholipid molecules in their environment. As a result, an increase in hydrocarbon mismatches takes place. A comparison of the mean thickness of the hydrophobic part of the membrane dL ¼ (dg þ df)/2dp gives dL < dp. We can thus expect an increase in the phase transition temperature Tc in the BR-DMPC system, and this is in agreement with the experimental results [158]. Thus, BR influences its lipid environment at large distances–over a region at least 12–20 nm in diameter. For the geometry and the range of deformation, not only is

Figure 3.37 Dependence of the phase transition temperature Tc of proteoliposomes of DMPC containing BR on the molar ratio of BR/DMPC (xp). (1) DSC experiment. Theoretical calculations: (2) x ¼ 16.2 A˚, N ¼ 1; (3) x ¼ 16.2 A˚, N ¼ 3; (4) x ¼ 10.48 A˚, ˚ , N ¼ 3 [51]. N ¼ 1; (5) x ¼ 10.48 A

the size of the hydrophobic mismatch important; determination of the characteristic decay length of the perturbation, which depends on the elastic parameters of the membrane, is important as well. In addition to native integral proteins, model proteins, e.g. a-helical peptides are also used. Using these model systems, it has been confirmed that a mismatch between the hydrophobic moiety of the protein and lipid bilayer results in changes in the thermodynamic properties of the lipid bilayer [86]. Experiments with simple peptides incorporated into lipid membranes of different hydrophobic thicknesses revealed that a long peptide can incorporate into either a thick or a thin bilayer, in the later case by tilting. In contrast, short peptides cannot incorporate into too thick a membrane and instead will form aggregates [142,159,160]. The possibility of peptide tilting in thin bilayers has also been demonstrated by the molecular dynamic simulation method [161,162].

Table 3.7 The results of the calculation of the minima of deformation energy Fmin, decay length x and x0 and the corresponding radial range of the distorted region (r  r0) Phase state

Fmin, kT

(r  r0)e, Å

n, Å

Fmin, kT

(r  r0), Å

n0 , Å

Gel Fluid

1.33 1.68

62.5 55

16.2 13.5

74 62.5

10.48 9.6

48.5 40

Index e denotes the exact solution of Equation (3.44) for the BR-DMPC system (see the text).

Biological Membranes and Membrane Mimics

While considerable attention in experimental and theoretical studies was focused on the problem of the interaction of integral proteins with lipids, less theoretical work is known for the analysis of the lateral organization of peripheral proteins. However, there is a considerable amount of literature focused on the organization of biopolymers and receptors on lipid monolayers [36,163]. This is particularly connected with the development of biosensors. Experimental aspects of the interaction of peripheral proteins with the membrane surface are reviewed in a paper by Kinnunen et al. [164]. This topic is also a rather attractive area in connection with adsorption of DNA onto the membrane surface, which can be an initial step for the subsequent translocation of this molecule into the cell, e.g. by means of electroporation. In these systems membrane elasticity plays an important role. It has been shown that the interaction of DNA and its complexes with cationic surfactants results in considerable changes in the BLM elasticity [133]. Changes in the E? of a supported BLM have also been observed during adsorption of model a-helical peptides onto the membrane surface [165]. Theoretical work focused on the analysis of the interaction mechanisms of peripheral proteins with the membrane surface has been reviewed in a paper by Gil et al. [140]. Various approaches in this analysis include, for example, molecular dynamic simulations of the association of peripheral proteins with fully hydrated lipid membranes. This method has been applied to study the interaction of phospholipase A2 (PA) with the membrane surface [166]. The authors obtained detailed information on PA conformation and also analyzed the enzymatic activity of this protein. Another approach was developed by Heimburg and Marsh [167] and is concerned with the expression of isotherms of binding for the adsorption of charged proteins (e.g. cytochrome c (cyt c)) to a charged surface (e.g. dioleoylphosphatidyglycerol bilayers). It has been found that the cross-sectional area of cyt c is equivalent to 12 lipids in a fluid bilayer and that the charge of the protein in a membrane is lower in comparison with the net charge of the native protein in solution. Currently one of the most effective approaches involves the application of Monte Carlo simulations [168]. This method was applied to the study of the aggregation of cyt c in a dimyristoylphosphatidylglycerol bilayer and revealed a high potential for the analysis of protein-induced phase separation in binary lipid mixtures, where protein prefers lipids of certain configuration. The binding of proteins to the membrane surface may result in changes in the surface potential of the BLM as well as changes in the dynamics of reorientation of the

127

dipole moments connected with the head groups of phospholipid. The fundamentals of membrane potentials and dipole relaxation will be considered below.

3.7 MEMBRANE POTENTIALS 3.7.1 Diffusion Potential Unmodified bilayer lipid membranes represent an insulating layer with very low permeability for ions or other charged molecules. However, a BLM can be modified by ionic channels or carriers, which allow the transport of charged species by diffusion, either by means of a gradient of concentration or of a gradient of electric potential or both (see Section 3.9). In the case of a potential gradient it is the so-called diffusion membrane potential, or Nernst potential, that is the driving force for ionic transport across the membrane. The Nernst potential is determined as the difference between the potential inside and outside the cell (or the inner and outer sides of the membrane) Dj ¼ jin  jout ¼ 

RT Cin ln zi F Cout

ð3:46Þ

where R is the gas constant, T is temperature, F is the Faraday constant, z is valence of ion i and Cin and Cout are the concentrations of ion i at the inner and outer sides of the membrane, respectively. For measurement of the diffusion potential of a BLM, electrometers with high input resistance should be used, but even a simple pH meter can be applied for this purpose, assuming that two reference electrodes (e.g. Ag/AgCl) are used.

3.7.2 Electrostatic Potentials In addition to the diffusion potential there exists a membrane potential between the polar part of the membrane and the bulk of the electrolyte, the so-called border or electrostatic potential. The electrostatic potential at the membrane–solution interface is composed of two major components stemming from surface charges and dipoles, respectively (see, for example, [51,169–173]). In the case of charged lipid molecules there is a diffuse ionic double layer potential or surface potential originating from a fixed charge layer in combination with ions from the adjacent aqueous electrolyte solution. Its maximal value relative to the bulk solution lies just at the interface and can be described approximately by either the Gouy–Chapman or the Stern model (see e.g. [48,168]). In this section we will use the Gouy–Chapman potential FGC. In the polar head group

128

Bioelectrochemistry

Coulomb interaction between the ions (see [169]). Using the appropriate boundary conditions to solve the integral equations leads to the Gouy– Chapman equation X s ¼ ½2ee0 RT ci ðwÞfexpð  zi FFGC =RTÞ  1g 1=2 i

ð3:48Þ

Figure 3.38 Schematic representation of electrostatic potentials at the membrane–solution interface. The fixed charge surface potential (Gouy–Chapman potential FGC) and surface dipole potential Fd are shown. Indices 1 and 2 refer to the two sides of the membrane. The total surface potential on each side of the membrane Fm is given by FGC þ Fd. The difference between the surface potentials on the two sides, DFm ¼ DFGC þ DFd is called the transmembrane potential or ‘internal’ potential. Ua is the externally applied dc voltage [51].

region there is a further potential jump, Fd, resulting from the molecular dipoles of the lipids themselves or from oriented water molecules. This is the so-called dipole potential. Further charge and dipole contributions may come from adsorbed species. The total surface potential, Fm, is given by the sum Fm ¼ FGC þ Fd. Experimentally, FGC and Fd can be distinguished by the dependence of FGC on the ionic strength. Figure 3.38 gives a schematic representation of the electrostatic potential across a bilayer. The difference in the heights of the two corners at zero applied voltage, DFm, is equal to the difference between the surface potentials of the two sides of the bilayer DFm ¼ DFGC þ DFd

ð3:47Þ

Therefore DFm is a measure of the asymmetry of the electrostatic potentials associated with the membrane. Surface potentials are either localized strictly within the surface region (Fd) or extend, at most, to a limited distance from it (FGC). At equilibrium it is not possible to make a direct measurement of these potentials with electrodes in the bulk phase except at zero ionic strength. Gouy–Chapman Potential and the Determination of Surface Charge Density s The electrostatic potential at a charged surface in contact with an electrolyte reflects both the surface charge density and the redistribution of ions in the electrolyte solution in the presence of this potential. A description of the system is based on the Boltzmann equation to describe the concentration of each ionic species as a function of the electrostatic potential, and the Poisson equation to describe the

where ci(o) is the bulk concentration of species ‘i’ with charge zi. Other symbols have their usual meaning. For the general case, with multivalent ions present, there is no explicit solution for FGC as a function of s, but if only univalent ions are present the Gouy–Chapman potential, adapted to convenient units and at T ¼ 25  C, is given by FGC ðmVÞ ¼ 50:8 ln ½s þ ðs2 þ 1Þ1=2

ð3:49Þ

pffiffi where s ¼ 1:36s= c and s is in elementary charges per 2 nm , and c denotes the concentration of the 1 : 1 electrolyte in mol l1. It is evident from this equation that the surface potential resulting from a given charge density depends on the ionic strength of the solution: higher ionic strengths are said to shield the surface charge, resulting in a lower surface potential. Although it is not possible to measure the surface charge density of a membrane directly, it is possible to determine s by measuring the change in the Gouy– Chapman potential upon a change of the ionic strength. For this purpose the membrane is usually formed at relatively low ionic strength (e.g. 10 mmol l1) and the transmembrane potential is measured. Subsequently the ionic strength on one side of the membrane is increased, e.g. to 110 mmol l1, by addition of a small amount of concentrated electrolyte, and the transmembrane potential is again determined. The change in transmembrane potential measured in such a shielding experiment is just DFGC, as the dipole potential Fd is insensitive to the ionic strength (see [174]). Curves showing FGC vs reciprocal surface charge density for two different ionic strengths, as well a plot of the difference between these curves, are shown in Figure 3.39. Given the measured DFGC for the given change on ionic strength, the surface charge density can be read directly from such a plot. If DFGC is known to an accuracy of about 1 mV, then the minimal surface charge density that can be detected in 10 mmol l1 electrolyte is about 1.4 · 103 elementary charges nm2 (see Figure 3.39, curve 1). The biomembranes are negatively charged due to the presence of negatively charged phospholipids, as well as due to the negative charge of the proteins. Many adsorbed polyelectrolytes also have isoelectric points below neutral pH [172]. The surface charge density of biomembranes is usually between 0.17 e nm2 (frog node) [175] to 2.3 e nm2 (crayfish) [176] (e is the elemen-

Biological Membranes and Membrane Mimics

129

The dipole potential of monolayers, i.e. the boundary potential between the hydrocarbon center of the membrane and the bulk aqueous phase is typically several hundreds of millivolts [176]. Surface Potentials and Diffusion Potentials

Figure 3.39 Relationship between the Gouy–Chapman fixed charge surface potential and the membrane surface charge density s (e nm2, where e is the elementary charge). For a given value of the membrane charge density the Gouy–Chapman fixed charge surface potential will be reduced when the ionic strength is increased (curve 1: 10 mmol l1, curve 2: 110 mmol l1). Concentrations refer to 1 : 1 salt. The dotted curve represents the difference between the curves at 10 mmol l1 and 110 mmol l1. Point A corresponds to the shielding experiment described in ref. [173]. (DUCmin ¼ 20.6 mV, s ¼ 8 · 102 e nm2) [51].

tary or electron charge: 1.602 · 1019 C). This is different to the model membranes, where the charge density can vary to a large extent and can be both positive or negative depending on the lipid used: 2.5 s 2.5 e nm2 [172]. A qualitative interpretation of Fd is generally given in terms of analogy with a capacitor: a polarized molecule with an effective dipole moment M ¼ qd is similar to two conducting phases separated by a distance d and enclosing a charge density q. The dipole moment is expressed as Debye (D). The potential difference DFd is given by DFd ¼ 4pqd=e

ð3:50Þ

where e is the dielectric constant. If there is an array of n dipoles per unit area (3.49) DFd ¼ 4pn M ? =e

ð3:51Þ

where M? is the normal component of the dipole moment to the surface. There are several contributions to DFd: (1) the change induced by reorientation of the water dipoles in the presence of the monolayer-forming molecules; (2) the dipoles of the monolayer-forming molecules, namely those of polar head groups of phospholipids and those of the alkyl parts, which can be located in different dielectric constant. Then DFd ¼ 4p

X ðn M ? =eÞ

ð3:52Þ

It is worth emphasizing the difference between the transmembrane potentials and the membrane potentials (‘diffusion potentials’) resulting from selective permeability mechanisms (see Section 3.9). In contrast to surface potentials, diffusion potentials are due to a concentration difference between the aqueous phases and are thus a bulk property that can be measured directly. Surface potentials and diffusion potentials are completely independent conceptually and easily distinguished experimentally. An example of a non-zero transmembrane difference in surface potential (DFm) for which no potential difference can be detected between the bulk phases, is presented in the paper by Schoch et al. [178], where a charged membrane, made permeable to monovalent cations by nonactin, is asymmetrically shielded by calcium ions. This results in a non-zero transmembrane potential DFm, yet no bulk potential difference, as evidenced by a current–voltage curve that passes through the origin. The opposite situation, with DFm ¼ 0 but a non-zero diffusion potential, would be found, for example, with a membrane made of neutral lipids in the presence of nonactin and a gradient of monovalent cations. With no current flowing in the external circuit a diffusion potential would establish itself across the membrane, i.e. the current–voltage curve for this system would pass through i ¼ 0 at a voltage equal to the diffusion potential. (The potential drop takes place across the membrane itself and the resulting electrostrictive force will cause an increase in membrane capacitance C to reduce C to its minimum value. The transmembrane potential, due in this case to the diffusion potential between the bulk phases, would have to be brought to zero by external circuitry. Thus, in this situation, a measurement of the membrane capacitance as a function of the applied voltage would show a minimum capacitance at Vappl ¼ 0). A case in which both DFm and diffusion potential are non-zero is, of course, also possible. 3.7.3 Methods of Surface Potential Measurement Measurements on Monolayers and Vesicles The dipole potential can be measured by various methods, such as the ionizing electrode method or the vibrating plate method [177]. TREK Inc. (USA) [179] produces high sensitive electrostatic voltmeters, e.g. model 320C, that in connection with electrode model 3250 can measure

130

Bioelectrochemistry

the surface potential with an accuracy of 1 mV. This sensitive electrode is electromechanically vibrated to produce capacitive modulation between the electrode and the test surface. If the voltage on the test surface is different to the voltage on the reference surface (probe housing), an ac signal is induced upon the electrode by virtue of this modulation in the presence of the electrostatic field. The amplitude and phase of this ac signal are related to the magnitude and polarity of the difference in potential between the test surface and the probe housing. The TREK electrostatic voltmeter can be directly connected to the electronic unit of a NIMA trough [46], so fully computer controlled measurement of the surface potential under compression of the monolayer can be performed. For micelles and vesicles other methods for estimation of the various membrane potentials using non-electrode techniques have been developed (e.g. [110,171,180]). These methods are based on utilizing molecular probes. These probes either redistribute or change their molecular properties in membrane-associated electric fields. Often they respond in both ways. Certainly, most of the fluorescent, radioactive, spin labeled NMR probes as well as charged collisional quenchers are derived from various ions, and hence are sensitive to the membrane potential (see [172] and references therein for more details). A rather popular method for studying the electrostatic membrane potential is to measure the so-called x potential. This is done by determining the mobility of charged lipid vesicles in an external electric field. A charged vesicle placed into the external electric field Eex begins to move due to the electrostatic force and drags part of the diffuse double layer with it. The x potential is defined as the electrostatic potential at the plane of shear between the membrane-associated and stationary part of the double layer. An analytical expression for x potential has been found for particles that do not interact with the solvent. When a membrane vesicle of a radius r is placed in an electric field Eex, it will move with constant speed v, because of the balance between electrical force and the resistance of the medium with viscosity Z. The value of the x potential can be then calculated according to the Smoluchovsky equation x ¼ 4phv=ðeEex Þ

ð3:53Þ

(see [172] for the theory). Usually x  Fm. The difference between the potentials x and Fm will decrease with decreasing potential gradient, dFm/dx. This difference decreases in diluted solutions. Apart from the experimental difficulty of producing a uniform population of spherical particles with unknown surface conductivity, which limits the

accuracy of laser Dopler experiments at least, the uncertainty associated with the thickness of the double layer dx, at which the potential is determined, is the main drawback of electrophoretic potential measurements. Typically, a value dx ¼ 0.2 nm is assumed. However, the value of dx clearly depends on the interfacial membrane structure and hydration, owing to the fact that the plane of shear lies in the region where specific lipid–water effects dominate. It is thus probable that the actual value of dx changes with the membrane or solvent composition, temperature etc. The experimentally observed increase in the membrane hydration at the main phase transition temperature of the lipids thus offers one explanation for the observed increase in x potential at such a phase transition which is inexplicable in terms of the simple Gouy–Chapman model (see [172] for more details). Measurements on Planar Bilayers Direct measurement of the surface potential of planar bilayers is not possible, as mentioned above. It is, however, possible to measure the difference of the surface potential, DFm, and this is the parameter determined by the methods reported below. By varying the ionic strength on each side of the membrane, the fixed charge surface potentials, and thus the surface charge density, can be determined for the two sides independently (see ‘Gouy–Chapman Potential and the Determination of Surface Charge Density s’). In contrast to this, changes in Fd cannot be assigned to either side. The first determinations of the surface potentials of BLMs were based on current –voltage curves measured in the presence of an ion carrier. A method based on the dependence of membrane capacitance on the transmembrane potential (electrostriction) (see Section 3.7), and therefore independent of any transport mechanisms, was introduced in the mid 1970s [178,181–183]. The advantages and limitations of various approaches will now be considered in detail. Current–Voltage Characteristics. A complete description of the I–V curves for carrier-mediated ion transport involves both the surface potential depicted in Figure 3.38 and the Born self-energy of the charged species in the hydrophobic interior of the bilayer [184]. The theoretical basis for such analysis was described in 1970 [185], and many groups have published work in this field (e.g. [186– 189]). All interpretation is based on a more or less appropriate model to describe the potential energy barrier to carrier transport. An exact fit of the measured I–V curves can be a laborious procedure, and the detailed shape of the curve can easily be influenced by further factors, such as bilayer compressibility [200]. Nevertheless the surface potentials determined in this way have generally been found to

Biological Membranes and Membrane Mimics

be consistent with the prediction of Gouy–Chapman theory. While the method is experimentally relatively simple, and can yield information on both FGC and Fd, it harbors several disadvantages in practice. (1) A single, well-defined transport mechanism must be present. This usually means that a carrier substance must be added, which may limit the kind of experiment possible. (2) The exact results depend on the model and correction factors applied. (3) Each determination requires a measurement of the I–V characteristic and subsequent numerical analysis. This makes it impractical for the continuous automatic monitoring of surface potentials at frequent intervals [173]. Techniques Based on Electrostriction. The techniques discussed in this section are based on the compression of BLMs by a transmembrane electric field. The maximal bilayer thickness, corresponding to minimal capacitance, is found when the potential difference across the core of the bilayer is zero. (This ‘core’ corresponds to the hydrophobic region of the BLM: given the low dielectric constant and relative thickness of this layer, it will provide the main contribution to the overall membrane capacitance). As this core or ‘internal’ potential depends on both the intrinsic surface potentials of the bilayer itself (fixed charge and dipole potentials) and the externally applied potential, a change in the surface potentials can be compensated by an equal but oppositely directed external field. The externally applied voltage needed to achieve minimal capacitance has been designated the ‘capacitance minimization potential’ (UCmin, [181]) and is given directly by U Cmin ¼  ðDFGC þ DFd Þ

131

around a holding potential, Uh, which was then adjusted to give equal measured C(U) values at Uh DU. This corresponds to a discrete sampling of the C(U) curve to determine the applied voltage at which the bilayer capacitance has its minimum value. Automatic monitoring of DFm was achieved with either analog feedback circuitry or a computer. The noise level of UCmin was found to be 1–2 mV ptp for solvent-free BLMs (apparatus time constant 3 s, see [191]). In 1978 Alvarez and Latorre [182] reported the measurement of BLM surface potentials of solvent-free BLMs. Membrane capacitance was measured by recording the transient charging current following a small step change in applied potential. To enhance the resolution, the charging current of a matched R-C model circuit was subtracted from the BLM charging current using analog circuitry. Signal-to-noise was improved by digital signal averaging techniques. With a voltage step of 10 mV, the authors were easily able to detect changes (of 0.01 %) in BLM capacitance for bilayers having a coefficient of electrostriction of 0.02 V2. As reported by the authors, the measurements may take up to 25 s per point (512 repetitions at 20 repetitions s1). With large voltage jumps the time per point could be reduced considerably, but this may not be desirable in many situations. A method based on determination of membrane potential using current harmonics was described in 1980 [183] and was based on the generation of current harmonics when a sinusoidal voltage is applied to the membrane [51]. When membrane potential U1 is present and ac voltage with an amplitude U0 and frequency f is applied to the BLM, then in addition to a third current harmonic with amplitude I3 and frequency 3f, a second current harmonic with amplitude I2 and frequency 2f is also generated. The surface potential can then be determined from the equation

ð3:54Þ

The accuracy with which UCmin can be determined depends on the amplitude of the membrane compressibility coefficient: the bigger this parameter the easier the measurement becomes. The first report on the use of electrostriction for the determination of BLM surface potentials was given in 1976 [181] and described in detail in [178,184]. Experimentally a small ac signal (e.g. 1 kHz, amplitude 10–25 mV) was superimposed on an adjustable bias voltage. A continuous measurement of BLM capacitance was provided by a rectification of the 90 component of the ac BLM current, thereby allowing operation even at relatively large background conductances. The bias voltage was varied symmetrically in small steps (e.g., DU ¼ 25 mV)

Fm ¼  U 1 þ U 0 I 2 =ð4I 3 Þ

ð3:55Þ

The membrane potential can be determined either by compensation of the amplitude of the second current harmonic by an external voltage [183] or by measurement of the amplitudes of both second and third current harmonics [51]. For measurements using second current harmonics an ac frequency of 1 kHz has generally been used [183,192]. The optimal amplitude is determined by the BLM compressibility, but lies in the range 10–50 mV. As a rule, with a apparatus time constant of 1 s, the noise level corresponds to 1 mV for membranes having a coefficient of electrostriction down to 103 V2 at a frequency of 1 kHz. The compressibility of most BLMs is considerably higher than this. Care must be taken when

132

Bioelectrochemistry

BLMs have a voltage-dependent conductivity, i.e., nonlinear current–voltage characteristics, as this can also lead to the generation of second current harmonics. In this case the use of a phase-sensitive detector rather than just a tuned amplifier could increase the range of application of the technique. 3.8 DIELECTRIC RELAXATION The dielectric relaxation method is based on analysis of the time course of changes in the capacitance following sudden changes in the voltage applied across the bilayer [193]. Using this method, one can obtain information on reorientation of molecular dipoles and cluster formation. A symmetric voltage-jump (V to þV) across a bilayer can orient naturally occurring dipoles. The magnitude and time course of these effects depend on the structure of the bilayer and the bulk phase as well as on changes in membrane capacitance. The capacitance of the bilayer depends on the dielectric constant e of the bilayer material, the membrane area A and thickness of the membrane, d (see Section 3.3). The electric field could affect all these parameters. Therefore each parameter must be considered separately. This has been performed in [51,193]. No correlation was found between the normalized dielectric relaxation parameters and membrane area, showing that the effects were indeed related to the bilayer rather than to the border region [193]. 3.8.1 The Basic Principles of the Measurement of Dielectric Relaxation The dielectric relaxation is determined by measuring the time course of the displacement current following a step change in potential. A detailed description of the construction and operation of the apparatus is given elsewhere [193]. Briefly, a positive voltage is applied to the electrodes at time t ¼ 0. This causes a large charging current (I0) to flow, which decays with a time constant of RsCm, (Rs ¼ solution þ electrode resistance, Cm ¼ membrane capacitance). In addition, there may be relaxation currents caused by voltage- or time-dependent changes in Cm, and a

dc component through Rm. At time t ¼ 0, a negative voltage is applied to an R-C analog circuit which models the parameters of the experimental system. In this way a current is generated that is equal in magnitude, but opposite in sign, to that generated by the membrane charging current and the dc component. The currents from both circuits are combined, resulting in a canceling of the charging peaks and dc currents. Any capacitive relaxation is assumed to have an exponential time course IðtÞ ffi

X

I i0 e  t=ti

ð3:56Þ

i

The amplitudes are only meaningful when normalized in some manner: as an initial trial, the values were expressed per unit area, of which the simplest measure is the membrane capacity at zero voltage Cð0Þ ¼ C0 m . Thus, it is convenient to present relaxation amplitude (Iir ) as fractional or percent changes in capacitance, for which the complete equation is

I ir  DCi =C0 m ¼ I i0 ti =ðV 0 C0 m Þ

ð3:57Þ

This is related to the ‘dielectric increment’ through the dielectric constant. The latter is, however, not known for all the conditions met. Therefore the phenomenological ‘relaxation amplitude’ is used for analysis [193]. The resolution of the apparatus allows the detection of DCi of about 1 pF with time constant t > 1 ms. All relaxation phenomena reported here are considerably slower than this, therefore no inaccuracy is introduced from this source. 3.8.2 Application of the Method of Dielectric Relaxation to BLMs and sBLMs The method of dielectric relaxation allows one to study dynamic properties of BLMs and sBLMs and binding of macromolecules to the membrane surface. The method of dielectric relaxation allows us to determine the character-

Table 3.8 Relaxation times for the reorientation of dipole moments of a BLM prepared from soya bean phosphatidylcholine (SBPC) and an sBLM prepared from SBPC on the tip of a Teflon coated stainless steel wire or solvent-free on a smooth gold surface covered by alkylthiol [75]. Membrane system

s1, ms

s2, ms

s3, ms

s4, ms

s5, ms

s6, ms

BLM in n-decane sBLM in n-decane (stainless steel wire) sBLM solvent-free (smooth gold surface)

6.9 1.9 5.9 0.4 40 4.0

18 6 13.6 2.0 290 100

27.6 3.6 1903 760

51.7 4.2

85 9

120 12

Biological Membranes and Membrane Mimics

Table 3.9 Capacitance relaxation components of BLMs of different composition and following the absorption of the avidin–GOx complex from both sides of the membrane at a final concentration of 30 mmol l  1: (1) BLM from crude ox brain extract (COB); (2) B-BLM: BLM from biotinylated COB; (3) B-BLM þ A-GOx: BLM from biotinylated COB modified by avidin-GOx complex. System

s1, ms

s2, ms

s3, ms

BLM B-BLM B-BLM þ A-GOx

5.3 1.0 5.0 0.3 16.5 1.0

– 26.1 5.0 26.5 1.0

– 115 27 505 16

(Reprinted from Bioelectrochem. Bioenerg., 42, Snejdarkova, Rehak, Babincova, Sargent and Hianik, Glucose minisensor base of self-assembled biotinylated phospholipid membrane on a solid support and its physical properties, 8. Copyright 1987, with permission from Elsevier.).

istic time for the reorientation of dipole moments of phospholipid head groups. Due to the domain structure of lipid bilayers and the differing size of the clusters, we can expect different collective movement of the reoriented dipole moments following the application of symmetrical voltage jumps to the membrane. The results of determination of relaxation times for various BLM and sBLM systems are presented in Table 3.8. We can see considerable differences of relaxation times between different membrane systems. While free-standing BLMs are characterized by two relaxation times, the dynamics of sBLMs formed on the tip of a stainless steel wire can be characterized by up to six relaxation times. The reason for the increase in the number of relaxation components can be due to physical and chemical adsorption of phospholipids on the metal support. This adsorption could result in a different degree of immobilization of the lipid molecules and thus can lead to the appearance of more, different relaxation times. A similar result is also seen for membranes formed on thin gold layers with alkylthiols. In this case, the number of relaxation components is lower. However, a further increase of the duration of relaxation times takes place. This provides evidence about the strong restriction of movements of dipole moments. The result is in agreement with the increased values of the elasticity modulus of the latter membrane system in comparison with that of sBLMs formed on wires. The physical origins for different relaxation times for BLMs have been analyzed by Sargent [193]. It was suggested that the fastest relaxation times (several ms to tens of ms) could correspond to small amplitude reorientations of individual dipoles about an axis lying in the plane

133

of the membrane, while times of about a hundred ms reflect a rotational reorientation of individual molecules. Slow relaxation components (several hundred ms to ms) probably indicate the reorientation movements of domains or clusters of dipoles in the membrane plane. For comparison, an NMR study by Davis [194] gave dipole correlation times of 1–5 ms for lecithin vesicles, which is consistent with the results obtained from conventional BLMs as used in our experiments. Relaxation times in the ms and ms ranges, obtained by various techniques on phospholipid bilayers, were recently reported by Laggner and Kriechbaum [195]. Dielectric relaxation experiments allow us to study the binding of enzymatic complexes on the membrane surface and confirm the strong binding of the avidin–GOx complex to biotinylated membranes. These experiments were performed on a free-standing BLM. We checked in a stepwise manner how dipole relaxation times of phospholipid head groups changed upon modification of the lipids and membranes. The results obtained are summarized in Table 3.9. In this experiment current relaxation curves were averaged and the standard deviation taken as the experimental uncertainty. Native BLMs formed from crude ox brain extract exhibited one relaxation time of 5 1 ms. Additional relaxation components 115 27 ms and 26 1 ms appeared in BLMs modified by biotin. Addition of the avidin–GOx complex to the electrolyte (final concentration 30 nM) on both sides of the biotinylated BLM resulted in the appearance of a slow component, 505 16 ms. The appearance of this slow component presumably represents a collective motion of coupled dipole moments and reflects clustering in the membrane, induced by binding of the avidin–GOx complex [196]. Considerable changes in relaxation times have also been observed when short oligonuclotides modified by palmitic acid were incorporated into BLMs and sBLMs as well as during hybridization with complementary oligonucleotide chains at the membrane surface [197]. Dielectric relaxation methods can also be very useful for studying the binding of various macromolecules on the membrane surface, such as short peptides [198] and local anesthetics [199]. 3.9 TRANSPORT THROUGH MEMBRANES An important function of biomembranes consists in regulation of the transport of ions or other molecules, e.g. nutrients into the cell, or waste products and toxic substances out of the cell. Thanks to the membranes, all cells maintain concentration gradients of various metabolites across plasma membranes and the membranes of cell organelles. Transport

134

Bioelectrochemistry

of species across the membrane can be performed by passive diffusion or by facilitated diffusion. In the latter case the compounds either diffuse via a channel-forming protein or are carried by a carrier protein [200]. 3.9.1 Passive Diffusion Passive diffusion is the simplest transport process. The driving force of this process is a concentration gradient of species across the membrane or a membrane potential gradient or both, i.e. species are transported by passive diffusion from regions of higher concentration to those of lower concentration, or from regions of higher potential to those of lower potential. The equilibrium condition is reached when the concentrations or potentials are equal at both sides of the membrane. Let us consider the transport of charged molecules, e.g. ions through a semi-permeable membrane, on the basis of electrochemical potential m. For diluted solutions (typically Ci < 0.1 mol l1) m ¼ m0 þ RT ln C i þ zi Fj

which is the Nernst–Planck equation. If electrically uncharged particles are transported through the membrane, or if dj/dx ¼ 0, then

ji ¼  uRT

dCi ¼  D grad C i dx

ð3:61Þ

ð3:58Þ

where m0 is standard chemical potential (i.e. the chemical potential of one mole of the species), Ci is the concentration of molecules or ions, j is the potential, other parameters have their usual meanings. The flux Ji of the ions across the membrane can be determined from the Teorrell equation

where D ¼ uRT is the diffusion coefficient expressed in m2 s1 Equation (3.61) is the expression for Fick's first law of diffusion. If the concentrations of species at two sides of M the membrane are CM 1 and C 2 , respectively (Figure 3.40)

Ji ¼  D

ji ¼  uC i d m~dx

ð3:59Þ

where u is the mobility of the ions expressed in m2 s1 V1. Substituting Equation (3.58) into Equation (3.59) we obtain ji ¼  uRT

Figure 3.40 Schematic representation of the ion flux across the membrane.

dCi dj  uCi zi F dx dx

ð3:60Þ

M CM 2  C1 d

ð3:62Þ

where d is the membrane thickness. Because it is difficult to M determine the concentrations CM 1 and C 2 , for practical purposes, the following equation is used ji ¼  Pi ðC2  C1 Þ

ð3:63Þ

Table 3.10 Permeability coefficients for polar solutes across bilayers and biomembranes. (Adapted from [4,200])). Permeability coefficient, cm s1

Membrane Phosphatidylcholine (egg) Phosphatidylserine Phosphatidylserine : cholesterol (1 : 1) Human erythrocytes Frog erythrocytes Dog erythrocytes

Naþ



Cl

H2 O

Glucose